2026 ECNP Congress

PS02-2049

Biomarkers of aging capture cognitive and immuno-metabolic signatures of childhood adversity across data modalities

Vera N Karlbauer — M. Godara, M. Bauer, J.J. Mohn, M. Rex-Haffner, T. Namendorf, F. Hoffmann, S. Entringer, C. Buss, S.M. Winter, C. Heim, E.B. Binder, D. Czamara · Max Planck Institute of Psychiatry, Genes and Environment, Munich, Germany

2026 ECNP Congress

Introduction

Biological aging can be accelerated by early-life adversity (ELA), potentially elevating risk for mental and physical disorders later in life [1]. However, most biomarkers of aging have been developed in adult populations and their biological significance in children is yet unclear. Moreover, biomarkers derived from different data modalities may capture distinct aspects of aging and vary in their sensitivity to ELA. We therefore compared 27 biomarkers of aging across five data modalities and examined their associations with adversity and developmental health outcomes in a deeply phenotyped paediatric cohort.

Methods

Analyses were conducted in the Kids2Health cohort [4] of Berlin-based children aged 3–12 years (N=161–415 depending on data modality and outcome), 54% of which had experienced ELA in the form of childhood maltreatment or war and migration trauma. Biomarkers included twelve blood-based epigenetic clocks and two epigenetic estimators of telomere length, the buccal epigenetic clock PedBE, telomere length, two blood transcriptomic clocks, the metabolic clock MetaboAge, and seven brain imaging clocks. We assessed deviations between biological and chronological age using median absolute error (MAE), examined Pearson correlations between all biomarkers, and tested for associations with ELA and SES, body composition, cognitive ability, brain volume, immune markers, sex steroids and psychopathology.

Results

Within-modality biomarker correlations were moderate to high (r=.20–.96 for epigenetic clocks), but much lower between modalities, suggesting that different biomarkers might measure different facets of aging. Biomarkers that were specifically trained to predict chronological age and included children in their training sample, such as epigenetic clocks Horvath Skin&Blood and PedBE, had the lowest deviations between chronological and biological age in Kids2Health (MAE=1.15–1.21 years), while the highest prediction errors were observed for algorithms trained exclusively in adults, such as MetaboAge or GrimAge (MAE=20.43–33.8 years). However, only these adult-trained biomarkers were also associated with ELA and low SES, with nominal associations observed for epigenetic GrimAge (also PCGrimAge and GrimAge2) and MetaboAge, and post-FDR effects for the ENIGMA neuroimaging clock (beta=.14–.40, p<.05). Additionally, adversity-affected biomarkers showed further associations with higher BMI and body fat mass, lower brain volume, lower IQ, elevated pro-inflammatory markers, altered sex steroid concentrations, and higher rates of mental health problems (beta=-.71–.36, p<.05). This cognitive and immuno-metabolic signature of accelerated aging was not shared by child-trained biomarkers that had displayed low MAEs, but had not shown ELA associations.

Discussion

Our findings reveal a dissociation between accurate age prediction and capture of health-relevant information: While adult-trained biomarkers such as GrimAge, MetaboAge, or ENIGMA brain age displayed low intercorrelations and showed limited accuracy in predicting chronological age in children, they were associated with ELA and converged on a shared signature of altered immunometabolism and brain development. Conversely, child-trained biomarkers predicted chronological age accurately, but were insensitive to adversity and its health correlates. These results challenge the assumption that prediction accuracy is the primary criterion for selecting aging biomarkers in paediatric research. Biomarkers such as GrimAge, MetaboAge, and ENIGMA brain age may instead reflect biological processes shaped by ELA decades before disease onset, underscoring their potential relevance in young populations.

References

  1. Bourassa KJ, Sbarra DA. Trauma, adversity, and biological aging: behavioral mechanisms relevant to treatment and theory. Transl Psychiatry. 2024 July 12;14(1):1–13.
  2. Belsky DW, Moffitt TE, Cohen AA, Corcoran DL, Levine ME, Prinz JA, et al. Eleven Telomere, Epigenetic Clock, and Biomarker-Composite Quantifications of Biological Aging: Do They Measure the Same Thing? American Journal of Epidemiology. 2018 June;187(6):1220–30.
  3. Robinson O, Lau CE, Joo S, Andrusaityte S, Borras E, De Prado-Bert P, et al. Associations of four biological age markers with child development: A multi-omic analysis in the European HELIX cohort. eLife. 2023 June 6;12:e85104.
  4. Entringer S, Godara M, Hoffmann F, Klawitter H, Winter SM, Binder EB, et al. Maltreatment exposure is associated with metabolomic disease risk scores in children. Brain, Behavior, and Immunity. 2025 Nov;130:106130.