Dietary restriction promotes neuronal resilience via ADIOL

  1. Ana Guijarro-Hernández
  2. Shinja Yoo
  3. George A. Lemieux
  4. Sena Komatsu
  5. Abdullah Q. Latiff
  6. Rishika R. Patil
  7. Kaveh Ashrafi

Reviewed by

Read the full article See related articles

Discuss this preprint

Start a discussion What are Sciety discussions?

Listed in

Log in to save this article

Abstract

The steroid hormone 5-androstene-3β,17β-diol (ADIOL) was discovered in humans nearly a century ago, yet its physiological roles remain poorly defined. Here, we show that fasting and caloric restriction, two forms of dietary restriction, induce transcriptional upregulation of genes encoding CYP11A1, CYP17A1, and 17β-hydroxysteroid dehydrogenase family enzymes, promoting ADIOL biosynthesis. ADIOL, in turn, acts on the nervous system to reduce levels of kynurenic acid, a neuroactive metabolite linked to cognitive decline and neurodegeneration. This effect requires NHR-91, the C. elegans homolog of estrogen receptor β, specifically in the RIM neuron, a key site of kynurenic acid production. Consistent with the known benefits of fasting and caloric restriction on healthspan, enhancing ADIOL signaling improves multiple healthspan indicators during aging. Conversely, animals deficient in ADIOL signaling exhibit reduced healthspan under normal conditions and in genetic models of caloric restriction, underscoring the functional significance of this pathway. Notably, ADIOL does not significantly impact lifespan, indicating that its healthspan benefits are not simply a byproduct of lifespan extension. Together, these findings establish a physiological role for ADIOL in mediating the neuroprotective and pro-healthspan effects of fasting and caloric restriction and suggest that boosting ADIOL signaling may help narrow the gap between lifespan and healthspan. This positions ADIOL as a promising mimetic of dietary restriction effects on healthspan that could be used as a therapeutic strategy for age-related neurodegenerative conditions.

GRAPHICAL ABSTRACT

Article activity feed

  1. PREreview

    This Zenodo record is a permanently preserved version of a PREreview. You can view the complete PREreview at https://prereview.org/reviews/17991397.

    Reviewer report

    Dietary restriction promotes neuronal resilience via ADIOL

    Ana Guijarro-Hernández1, Shinja Yoo1, George A. Lemieux1, Sena Komatsu1, Abdullah Q. Latiff1, Rishika R. Patil1, and Kaveh Ashrafi1

    1 Department of Physiology, University of California, San Francisco, San Francisco, CA 94143, USA

    Summary

    This study identifies a physiological role of the steroid hormone ADIOL in mediating the beneficial effects of dietary restriction. Fasting and caloric restriction upregulate enzymes involved in ADIOL biosynthesis, and elevated ADIOL acts in the nervous system to lower kynurenic acid levels via NHR-91 signaling specifically in the RIM neuron. Enhancing ADIOL signaling improves several healthspan markers during aging, whereas loss of ADIOL signaling reduces healthspan even under normal or caloric-restriction–mimetic conditions. These effects occur without altering lifespan, indicating that ADIOL selectively enhances healthspan. Overall, the work positions ADIOL as a key mediator of the neuroprotective and pro-healthspan effects of dietary restriction and a potential therapeutic route for age-related neurodegeneration.

    The authors previously demonstrated that ADIOL promotes learning by reducing kynurenic acid levels (DOI: 10.1101/gad.350745.123). Here, they extend these findings to dietary restriction-mediated benefits. Given that ADIOL is conserved between worms and humans, and that dietary restriction exerts beneficial effects across species, the findings may have broader relevance to human health. While the manuscript presents interesting observations, such as ADIOL pathway-dependent increase in pumping rate upon fasting, additional experiments and clarification are required to fully support the authors' claims.

    Major comments

    1. On page 6 (3rd and 4th paragraph), the authors interpret the suppressed fasting-induced pumping increase in cyp-44A1 (CYP11A1), cyp-13A4 (CYP17A1), and F12E12.11 (17β-HSD) mutants as evidence that fasting enhances ADIOL production through upregulation of these genes. However, in Fig. 2b, cyp-13A4 expression is decreased upon fasting, F12E12.11 shows no significant change, and cyp-44A1 is only marginally upregulated. These data do not convincingly support their proposed model that fasting enhances ADIOL production through upregulation of these genes. To address this discrepancy, it might be helpful for the authors consider examining the expression or functional requirement of other ADIOL biosynthetic genes that are clearly upregulated by fasting.

    2. On page 8 (4th and 5th paragraphs), the authors conclude that ADIOL promotes healthspan based on enhanced pumping and thrashing during aging. However, ADIOL treatment also increases these activities in Day 1 adults, which are already at their peak physiological state in Fig 3b-e. The magnitude of improvement appears comparable between Day 1 and Day 5 animals. This raises the possibility that ADIOL broadly enhances neuromuscular performance independent of aging or healthspan, and it might not support their claim that ADIOL promotes healthspan. It would be helpful for the authors to clarify whether the effects are aging-specific or reflect general neuromuscular stimulation in the text.

    3. The title states that dietary restriction promotes neuronal resilience via ADIOL, yet the neuronal evidence that supports the central claim in the title is limited to the data that nhr-91 expression in RIM neurons rescues the reduced pumping rate in Fig 1c and 4c. The study does not test ADIOL's effects on learning or neuronal resilience per se. Although mxl-3 and daf-2 mutants were tested for learning, these genes influence numerous pathways beyond dietary restriction, making the interpretation less specific. Moreover, pumping, thrashing, and locomotion can reflect muscular rather than neuronal effects. A more specific title reflecting the presented data may be considered (e.g., "The ADIOL-kynurenic acid pathway mediates beneficial effects of dietary restriction.")

    Minor comments

    1. In Graphical Abstract, the arrows from CHOL to ADIOL are confusing, as PREG and DHEA might be seen as both receiving inputs and providing outputs. To improve interpretability, the authors can modify the curved arrows to direct ones between the molecules.

    2. In Abstract, it might be necessary to stress that the effects of kynurenic acid are known in humans to clarify what is known in C. elegans vs humans. The sentence "kynurenic acid, a neuroactive metabolite linked to cognitive decline and neurodegeneration." could be "kynurenic acid, a neuroactive metabolite linked to cognitive decline and neurodegeneration in humans."

    3. In Abstract (last sentence: "This positions ADIOL as a promising mimetic of dietary restriction effects on healthspan that could be used as a therapeutic strategy for age-related neurodegenerative conditions."), steroid hormones can have both positive and negative effects. Thus, it would be more informative if the authors discuss if there are any negative consequences of elevating ADIOL in worms in the Results or Discussion section.

    4. F12E12.11 has a gene name of nta-1. Please consider using the updated name of the gene to improve reader understanding.

    5. In Fig 1c, cex-1p::nhr-91 was used to express nhr-91 in RIM specifically. However, given the reported nonspecific posterior intestinal expression associated with unc-54 3′UTR used in their constructs (Thomas Boulin, John F. Etchberger, and Oliver Hobert, 2006, WormBook), it would strengthen the author's conclusion that RIM neurons mediate the effects of ADIOL if they directly demonstrated that cex-1p::nhr-91 expression is restricted to RIM and not present in other cells.

    6. In Fig 1c and Fig 4c, Different nhr-91 constructs were used for expressing nhr-91 in RIM (cex-1p::nhr-91) and RIC (tbh-1p::nhr-91cDNA::SL2::GFP) (Supplementary Table 1). This difference is not discussed, and the functionality of the nhr-91cDNA::sl2::GFP transgene is not demonstrated. It remains possible that a nonfunctional transgene in RIC fails to rescue the phenotype. It would be more supportive if the authors demonstrate the transgene functionality with an additional experiment. If the nhr-91cDNA::sl2::GFP transgene is functional, the rescue experiment would further support that nhr-91 does not act in RIC.

    7. In page 5 (1st paragraph), the authors refer to mxl-3 mutants as a caloric restriction model because starvation represses mxl-3 expression. However, starvation and caloric restriction are distinct interventions, and physiological responses are different depending on caloric restriction between starvation (e.g. starvation can induce developmental arrest as a stress response. To improve reader understanding, the authors should consider clarifying this distinction in this paragraph.

    8. For clarity, it would be helpful if the authors could define several acronyms that are used throughout the text, including HSD, F17, and KAT.

    9. In page 5, A brief explanation of the behavioral learning assay used would improve clarity, as multiple learning paradigms exist.

    10. In Fig 2b, consider adding a legend explaining what the numerical values represent.

    11. In Fig 2c, the use of 2–dCt rather than fold change, which is used in Fig 2d,e. It might be helpful to explain why the authors used different analysis methods for clarity.

    12. In page 7 (2nd paragraph) and Extended Fig 2a, the Day 1 pumping rates are different for the three genotypes, and this result is not consistent with the data in Fig 1a-c. The authors might need to explain this discrepancy for clarification.

    13. On page 8 (1st paragraph), the authors did not explain what FUDR is or why assays were conducted without FUDR to better assess age-related effects. Please consider explaining the rationale of using FUDR in the experiments to improve reader understanding.

    Competing interests

    The authors declare that they have no competing interests.

    Use of Artificial Intelligence (AI)

    The authors declare that they did not use generative AI to come up with new ideas for their review.

  2. PREreview

    This Zenodo record is a permanently preserved version of a PREreview. You can view the complete PREreview at https://prereview.org/reviews/17991795.

     In this manuscript, the authors propose that the pro-healthspan effects of dietary restriction (DR) are mediated by the steroid hormone ADIOL. They conclude that in the nematode C. elegans, DR induces intestinal ADIOL production, which then acts neuronally to promote healthspan. Importantly, the authors make a distinction between healthspan and lifespan; ADIOL improves only health measures but does not alter the longevity of C. elegans. They additionally suggest that ADIOL acts mechanistically through a downstream kynurenic acid (KynA) signal to mediate these pro-healthspan effects and that some pro-healthspan effects of dietary restriction (DR) can be attributed to this ADIOL pathway.

    The effects of ADIOL on healthspan are convincing, and there is strong evidence presented that this effect is indeed decoupled from lifespan: ADIOL enhances multiple measures of fitness without extending lifespan and additionally does not seem to be the molecule responsible for longevity effects in long-lived mutants. The paper would be improved if the authors could clarify and/or further substantiate three of their main claims, as explained below: 1. KynA is the downstream effector of ADIOL.  2. ADIOL is a neuroprotective/anti-aging signal. 3. DR induces upregulation of ADIOL signaling. Clarifying these issues would provide a more complete and more precise mechanistic understanding of the entire pathway behind DR's leading to enhanced healthspan, which would be a more impactful statement of the paper.

    Major points:

    1. A central explicit statement in the paper is that KynA is the (neuronal) effector metabolite that ADIOL induces to mediate its healthspan effects (Abstract/line 6, Page 4/line 4, Page 10/line 31, and the text/figure titles of Fig 3 and 4). The study utilizes mutants from a previous publication of the authors (Lemieux,.. Ashrafi, 2023, PMID: 38092521) defective in key genes that act directly in the KynA pathway, kmo-1, nkat-1 and aat-1 (Figure 4d-e). It would greatly strengthen the authors' claims about KynA's being downstream of ADIOL if they were to include those mutants in more of their analyses / epistasis analyses. In particular, extending their inclusion of kmo-1 and nkat-1 mutants to the aging healthspan assays of learning and pharyngeal pumping in Fig. 3a-b and testing double mutants of nkat-1 and nhr-91 in healthspan assays would clarify the relative contribution of ADIOL on the KynA pathway in an aging context and distinguish the findings described in this manuscript from their KynA investigations reported in earlier papers. This issue is especially pertinent as the authors state in the discussion that "KMO-1 activity may influence KynA production independently of ADIOL", indicating that it is possible for KynA to not be directly downstream of or influenced by ADIOL.

    2. The authors use the term "neuronal resilience" (in the title) and "neuroprotective" (in the abstract) to describe ADIOL's effects on C. elegans physiology. Such terms seem to imply that ADIOL protects neurons against a decline of the functional measures of learning, pharyngeal pumping and motility.

    i) Though nhr-91 mutants show an earlier decline in healthspan as measured by the learning assay (Fig. 3a), all other aging assays and experiments (Fig. 3-4) demonstrate that ADIOL addition alone at day 1 already induces enhanced measures of pro-healthspan attributes. The authors should consider more precisely defining/characterizing the effects of ADIOL; it does not just prevent or protect against age-related decline but also actively raises the baseline measures of healthspan. One way to disentangle these issues would be to test earlier-aged worms, for example at the L4 stage, to determine if there is an age at which ADIOL does not simply increase the healthspan baseline.

    ii) Relatedly, from a loss of function direction for ADIOL, how do nhr-131 mutants perform in the learning assay in Fig. 3a at earlier stages, such as L4? Are they healthspan defective in general or do they have an accelerated aging-related decline that happens earlier than day 1?

    iii) As the authors were also able to measure ADIOL and KynA levels in previous publications (e.g. Ref 9), direct measurements of ADIOL and KynA levels across aging timepoints should be possible and also would help bolster a direct connection of ADIOL to aging.

    iv) Furthermore, all these measures are a function of both neuronal and muscular contributions; ADIOL has a likely intestinal origin and thus likely non-cell autonomous effects, and there is insufficient evidence to rule out an effect of muscle contribution here (see also minor point 1 about ADIOL's site of action).

    v) If the authors were willing to rephrase, they could consider a straightforward term to use: "pro/enhanced healthspan" instead of "resilience" or "protective".

    3. Fig. 2 shows expression data of ADIOL synthesis genes to support the claim that DR activates ADIOL signaling. Though there is a general trend across the aggregated data that supports the stated conclusions, some of these observations do not follow the general trend and are not as clear cut. In particular, the model in Fig. 2f attributes the effects of 2h-fast, and daf-2 and mxl-3 mutations to particular enzymatic reactions, but the data do not follow as cleanly. For example, if the enzymatic reactions are linearly arranged as shown, the nhr-131 mutation would be expected to show strong effects that block upstream processing of ADIOL; accordingly, we would not see an increase of cyp-33C5 expression in the nhr-131 background in Fig. 2b-c in the 2h fast condition. The authors do recognize that the results are not clear cut with phrases such as "some cases (Page 6 line 16)" and "partial requirements likely reflect enzymatic redundancy" (Page 6 line 23). However, given the claim as written in the title of the section is: "ADIOL levels increase under fasting and CR conditions through the upregulation of CYP11A1, CYP17A1 and 17-βHSDs orthologs (Page 5 line 25)", the authors either should consider strengthening their claim by direct measurements of ADIOL (and KynA) in DR conditions or rephrasing the wording they use in their section title etc..

    Minor points:

    1. The authors specify that the effects of ADIOL-enhanced healthspan are neuronally mediated. I understand that their previous work has shown neuronal contributions to the effects of KynA on healthspan (Lemieux,.. Ashrafi, 2023). However, it would help readers of this paper if such prior evidence were clearly and precisely spelt out in the appropriate areas; in particular, I refer to the expression patterns of ADIOL-related genes and gene promoters used for rescue experiments. Both the cex-1 and tbh-1-based promoters that are used to show RIM and RIC expression have non-neural expression, so it is important to be precise about what is demonstrated and what is proposed about ADIOL's site of action and about the difference between necessity and sufficiency in wording. In particular, the two pieces of data used to show a RIM-specific site of action for ADIOL in aging provide relatively weak evidence: (1) in Fig.3d the magnitude of rescue of nhr-91 in RIM for pharyngeal pumping is not particularly striking, and (2) in Fig. 4c the baseline rate of thrashing is already higher in nhr-91 mutants at day 5 compared to wild-type and is in fact an opposite result to what is expected of an ADIOL-insensitive phenotype. More context about ADIOL-gene expression data would also be helpful, i.e. where are nhr-91 and nhr-131 expressed?

    2.     Given the major point about ADIOL's ability to enhance baseline measures of healthspan, in Fig, 1b, it seems relevant to assess what ADIOL addition does to well-fed WT worms. That would be informative, as the combination of exogenous ADIOL and 2h fast might already be hitting a ceiling effect.

    3.     The mxl-3 mutant has a striking learning assay phenotype, but its enhanced lifespan phenotype seems really weak. In fact, it seems that the median survival of the mxl-3 mutant (Fig 5e) here is much lower compared to reported values. (17 days (Fig 5e) vs ~28 days (Ref 43)). It would be helpful for the authors to clarify any potential reasons for their thoughts about this discrepancy.

    4.     In the molecular pathways depicted in the graphical abstract and figures, molecules like hormones (ADIOL) are not distinguished stylistically from proteins like NHR-31 and NHR-131. Furthermore, the same colors are used for both PREG and NHR-91 as well as CHOL and KynA, even though these are different molecules acting in different parts of the pathway. The authors might want to consider marking the differences in these pathway nodes by changing styles and colors to enhance readability.

    Competing interests

    The authors declare that they have no competing interests.

    Use of Artificial Intelligence (AI)

    The authors declare that they did not use generative AI to come up with new ideas for their review.

  3. Version published to 10.1101/2025.06.25.661551 on bioRxiv