Loading evaluation data…
Abstract Drugs for aging-related diseases may modulate aging itself, but standard clinical trial designs cannot detect such effects. Aging clocks could close this gap, but epigenetic models often yield inconsistent, hard-to-interpret results. In contrast, proteomic clocks, by tracking the immediate effectors of biological change, may excel in providing aging biomarkers or mechanistic insight. Here we compare six proteomic clocks (ProtAge, OrganAge mortality , OrganAge chrono , PAC, ipfP3GPT and PAOPAC) on serum proteomes from a published 12-week phase 2a trial of the candidate anti-fibrotic drug rentosertib in idiopathic pulmonary fibrosis. We measure the variance between the clocks and find that all six clocks consistently predicted lower biological age in treated arms. However, proteomic clocks alone cannot deconvolute aging- and disease-specific effects. We addressed this issue indirectly through pathway analyses that identified potential anti-aging shifts in senescence and metabolic processes alongside the anti-fibrotic activity of rentosertib. This work supports the goal of dual-purpose clinical trial designs that integrate aging endpoints into studies for specific disease indications.
all six clocks consistently predicted lower biological age in treated arms
directionally convergent clock outputs from a small selected cohort, with significance concentrated by regimen and timepoint
proteomic clocks alone cannot deconvolute aging- and disease-specific effects
the IPF trial design cannot directly separate disease improvement from modulation of aging biology
combined proteomic aging clock and senescence marker analysis revealed potential geroprotective effects
exploratory computational analyses in a small cohort provide indirect rather than confirmatory geroprotection evidence
rentosertib’s effects on biological age cannot be attributed fully to its anti-fibrotic activity
FVC dissociation and external aging trajectories are suggestive but do not identify an independent aging effect
Derived from the full evaluation — not a separate score.
Strengths
Six independently developed proteomic clocks were applied longitudinally to treatment and placebo arms, providing a controlled setting for assessing agreement and treatment responsiveness.
↳ Results, Aging clock performance and Rentosertib-induced aging clock predictions; Figures 2–3
The analysis applies Benjamini–Hochberg correction and tests the aggregate clock signal using 100,000 patient-label permutations, with a sensitivity analysis excluding higher-grade adverse events.
↳ Methods, Statistics and reproducibility; Results, Rentosertib-induced aging clock predictions; Extended Data Figure 2
The paper explicitly states that proteomic clocks cannot distinguish aging modulation from disease improvement and limits its conclusion to an early, hypothesis-generating indication.
↳ Abstract; Results, Isolating geroprotective effects; Discussion
Limitations
Only 42 of 71 randomized participants entered the analysis after trial completion, consent, and complete-sample requirements. No reported comparison of included and excluded participants establishes that this subset preserved the parent trial's balance.
↳ Results, opening paragraph; Methods, Study participants
FVC dissociation, UK Biobank trajectories, and pathway analyses are suggestive but cannot distinguish direct aging modulation from IPF-related proteomic improvement. The paper acknowledges that confirmation requires non-IPF populations and direct validation.
↳ Results, Isolating geroprotective effects; Discussion
The Results and Methods identify NCT05938920, while reference 34 links to NCT05975983. This prevents unambiguous verification of the parent protocol until corrected.
↳ Results, opening paragraph; Methods, Trial design; Reference 34
The score reflects a genuine methodological advance: six clocks were compared longitudinally within randomized parent-trial infrastructure and supplemented by pathway, protein-trajectory, and UK Biobank analyses. Methodological Rigour is held in the lower band because the analyzed cohort comprises 42 completers who consented after the trial, without a reported included-versus-excluded comparison, and because no analysis isolates direct aging modulation. Reporting and context score more strongly because the Abstract and Discussion repeatedly acknowledge that geroprotection remains an indirect hypothesis. The inconsistent trial identifiers in the Methods and reference list must be resolved before the parent protocol can be traced reliably.
Nabu’s assessment, alongside the field’s view.
Are you an author of this paper?
Sound3.4
Confidence mediumThe study provides a head-to-head application of six proteomic clocks within a controlled clinical-trial setting and integrates clock outputs with pathway and population-trajectory analyses. Its advance is methodological rather than a definitive demonstration of geroprotection.
“no previous study has systematically compared several proteomic aging clocks in a clinical intervention”
The parent trial was randomized and blinded, and the ancillary analysis uses multiple-testing correction, mixed models, and permutation testing. Reliability is reduced because only 42 completers and consenters were analyzed and the indirect analyses cannot isolate aging modulation from disease modification.
“Full disentanglement of the two signals is not achievable within an IPF cohort”
The paper clearly distinguishes measured clock shifts from confirmed geroprotection and repeatedly traces uncertainty into its conclusions. The abstract's claim of consistency is somewhat broader than the regimen- and timepoint-specific results but does not reverse the paper's qualified interpretation.
“proteomic clocks alone cannot deconvolute aging- and disease-specific effects”
The paper situates proteomic clocks against epigenetic-clock limitations, prior intervention studies, and regulatory biomarker pathways. It explicitly identifies the evidence needed for confirmation, including non-IPF populations, multimodal profiling, and functional outcomes.
“Definitive confirmation, however, will require studies in non-IPF populations”
Concerns4 of 4 checks
Most cohort and analysis numbers are internally coherent, but the parent trial cannot be identified unambiguously because two different registration numbers are reported. A smaller discrepancy also appears in the number of participating locations.
Ethics approval, consent, competing interests, data availability, and code availability are declared. Registration of the parent trial is stated, but the inconsistent identifier makes the corresponding record ambiguous, and no preregistration is stated for the ancillary clock analysis.
Flags: 4 declared / 5 total
17 of 17 checkable references verified
17 references in manuscript
No retraction notice found in Retraction Watch.
Sources: Retraction Watch ✓
Where this paper’s evidence sits on the path from initial observation to real-world use.
The work demonstrates feasibility in a clinical setting but remains an exploratory ancillary analysis rather than a qualified biomarker or treatment-guiding result. Direct experimental validation and prospective functional endpoints are still required.
“The evidence described here provides an early indication”
AI-generated, human-governed. Something look off? Contact us to request a review.