Open annotations (there are currently
Altmetric provides a collated score for online attention across various platforms and media.
Computer prediction and genetic analysis identifies retinoic acid modulation as a driver of conserved longevity pathways in genetically diverse
https://doi.org/10.7554/eLife.104375.3
study explores the power of computational methods to predict lifespan-extending small molecules, demonstrating that while these methods significantly increase hit rates, experimental validation remains essential. The study uses all-trans retinoic acid in
as a model, providing genetic and transcriptomic insights into its longevity effects. The data are
in describing a robust, computationally informed screening process for discovering compounds that extend lifespan in this species.
https://doi.org/10.7554/eLife.104375.3.sa0
: Findings that have theoretical or practical implications beyond a single subfield
: Evidence that features methods, data and analyses more rigorous than the current state-of-the-art
During the peer-review process the editor and reviewers write an eLife Assessment that summarises the significance of the findings reported in the article (on a scale ranging from landmark to useful) and the strength of the evidence (on a scale ranging from exceptional to inadequate).
Learn more about eLife Assessments
Discovery of new compounds that ameliorate the negative health impacts of aging promises to be of tremendous benefit across a number of age-based comorbidities. One method to prioritize a testable subset of the nearly infinite universe of potential compounds is to use computational prediction of their likely anti-aging capacity. Here, we present a survey of longevity effects for 16 compounds suggested by a previously published computational prediction set, capitalizing upon the comprehensive, multi-species approach utilized by the
Intervention Testing Program. While 11 compounds (aldosterone, arecoline, bortezomib, dasatinib, decitabine, dexamethasone, erlotinib, everolimus, gefitinib, temsirolimus, and thalidomide) either had no effect on median lifespan or were toxic, 5 compounds (all-trans retinoic acid, berberine, fisetin, propranolol, and ritonavir) extended lifespan in
. These computer predictions yield a remarkable positive hit rate of 30%. Deeper genetic characterization of the longevity effects of one of the most efficacious compounds, the endogenous signaling ligand all-trans retinoic acid (atRA, designated tretinoin in medical products), demonstrated a requirement for the regulatory kinases AKT-1 and AKT-2. While the canonical Akt-target FOXO/DAF-16 was largely dispensable, other conserved Akt-targets (Nrf2/SKN-1 and HSF1/HSF-1), as well as the conserved catalytic subunit of AMPK AAK-2, were all necessary for longevity extension by atRA. Our results highlight the potential of combining computational prediction of longevity interventions with the power of nematode functional genetics and underscore that the manipulation of a conserved metabolic regulatory circuit by co-opting endogenous signaling molecules is a powerful approach for discovering aging interventions.
Aging is a primary risk factor for a myriad of chronic illnesses, health declines, and mortality. A central premise of research in the current aging field is that aging per se can be treated directly, leading to ancillary benefits across a broad range of age-related comorbidities (the ‘geroscience hypothesis’;
). But how best to identify compounds holding the potential for broad-spectrum effects across an individual’s lifespan? While comprehensive screens using model organisms such as the nematode
), a complementary alternative is to use emerging databases of compound-specific physiological effects to predict which compounds are most likely to lead to positive effects on extending lifespan (
). An advantage of this approach is that the predictive models should become better and better as the training set of positive hits continues to expand over time (
). Still, the efficacy of any predictive model is strongly dependent on the quality of the input data, and the well-documented heterogeneity of aging as a phenotype, as well as general challenges in reproducibility per se, create barriers to the successful application of predictive approaches to aging research. The
Intervention Testing Program (CITP) tests compounds for lifespan and healthspan effects across a genetic diversity panel of
of response across genetic backgrounds, the CITP has painstakingly focused on
across laboratories and trials via standardization of methods and a hierarchical statistical approach that accounts for experimental variation at a variety of levels of replication. These features make the CITP an ideal framework for testing computer predictions of longevity interventions and serve as the foundation for data collection for improved models in the future.
As a first step toward testing the efficacy of computational prediction of lifespan-extending compounds, we used a previously published set of compound predictions developed via an analysis of the overlap of drug-induced and aging-related gene expression and protein interactions (
) to develop a list of candidate compounds for further investigation using the CITP workflow. We prioritized compounds with the highest predictive scores and eliminated several compounds whose effects in
were already well characterized. Our analysis led to a set of 16 compounds (aldosterone, all-trans retinoic acid (atRA), arecoline, berberine, bortezomib, dasatinib, decitabine, dexamethasone, erlotinib, everolimus, fisetin, gefitinib, propranolol, ritonavir, temsirolimus, and thalidomide) selected for further testing. As outlined below, we found that of the five candidate compounds – atRA, berberine, fisetin, ritonavir, and propranolol – that extended median lifespan, propranolol and atRA conferred the largest positive effects. Potential confounding interactions of propranolol with the bacterial food of the nematodes led us to focus on atRA for more in-depth genetic and functional analysis.
atRA is an FDA-approved intervention used topically in dermatology and systemically as a chemotherapeutic adjuvant (
). Endogenously, atRA is the most bioactive retinoid derived from vitamin A, known to function as a highly conserved signaling ligand involved in transcriptional regulation (
, the presence of vitamin A metabolism pathway genes (
), metabolism of exogenous vitamin A into retinal and atRA (
fatty acid- and retinol-binding proteins for retinoids (
), and endogenous atRA detection in untreated animals (
) combine to suggest the presence of an endogenous nematode atRA signaling pathway. While conservation of the ligand atRA is well supported, the canonical vertebrate downstream retinoid receptors (RXR and RAR) that effect transcriptional responses have not been identified in nematodes. In contrast with the elusive retinoic acid receptors, however, the mammalian kinases modulated by atRA have extensively studied
orthologs. In humans, atRA modulates transcription via PI3K/Akt (
) kinase signaling. Functionally, kinase signaling is likely mediated by atRA regulation of the kinase phosphorylation state, as has been shown for Akt in mammalian (
Building upon our general screening approach, we present a more comprehensive genetic analysis of atRA impact on longevity that suggests functional conservation of atRA kinase regulation, as the effects of atRA on longevity require kinases encoded by both
and mammals, Akt kinases regulate powerful aging pathways (e.g., insulin-like signaling (IIS), FOXO, and Nrf2). Our genetic analysis of atRA longevity in
suggests that the FOXO/DAF-16 transcription factor is not necessary, consistent with atRA acting downstream of, or in parallel to, FOXO. In contrast to FOXO/DAF-16, the Akt-phosphorylation targeted Nrf2 homolog SKN-1 and heat shock transcription factor 1 homolog HSF-1, along with the co