Aging Monkeys Reveal Two Distinct Molecular Paths of Brain Aging

Aging remains the single greatest risk factor for Alzheimer’s disease and other neurodegenerative conditions, yet scientists still lack a complete molecular map that separates healthy brain aging from the earliest steps of pathological decline. A new study published in GeroScience takes a major step toward that goal by profiling proteins in the blood and cerebrospinal fluid of 66 rhesus macaques, animals whose biology, lifespan trajectory, and susceptibility to age-related disease closely mirror our own. The research, led by Giovanne B. Diniz and John H. Morrison at the University of California, Davis, together with collaborators at Stanford University, Wake Forest University, and the Icahn School of Medicine at Mount Sinai, reveals that aging unfolds along strikingly different molecular routes in the periphery and the brain, and that the oldest animals split into distinct proteomic subtypes that chronological age alone cannot explain.

The team used NULISAseq, a next-generation sequencing-based targeted proteomic platform developed by Alamar Biosciences that offers attomolar sensitivity and broad dynamic range while requiring only tiny sample volumes. The CNS Disease Panel 120 quantified roughly 130 protein targets spanning neuronal, glial, vascular, immune, and metabolic pathways. The macaques, aged between 5.97 and 33.58 years, were stratified into adult, pre-geriatric, and geriatric groups corresponding broadly to human adulthood, late midlife, and advanced age. Twelve human cerebrospinal fluid samples from the Stanford Alzheimer’s Disease Research Center, including aged controls, APOE4 carriers, and Alzheimer’s patients, were run in parallel to anchor the cross-species comparison.

The first headline finding is an asymmetry between compartments. In plasma, nearly 38 percent of measured proteins showed significant monotonic associations with chronological age, and the differences were overwhelmingly positive, meaning protein concentrations rose with age, consistent with increased production or impaired clearance. In cerebrospinal fluid, by contrast, only about 19 percent of proteins tracked with age, and the changes were more bidirectional, with some proteins rising and others falling. Effect sizes were substantial in both fluids, with upper-quartile eta-squared values approaching 0.5, indicating that age explains a large share of the variance for a subset of analytes. Principal component analysis reinforced the split: the dominant axis of plasma variance was significantly associated with age, sex, birthplace, and body weight, whereas the dominant cerebrospinal fluid axis was not tied to age at all, with age-related variance distributed across higher-order components.

Encouragingly for translational research, the canonical markers of neurosenescence behaved in macaques much as they do in humans. Neurofilament light and heavy chains, well-established indicators of axonal injury, rose robustly with age in both fluids, and levels in older macaques overlapped substantially with human reference values. The astrocytic proteins GFAP and CHI3L1, also known as YKL-40, increased with age in both compartments. Microglial markers told a more compartmentalized story: CHIT1 and TREM2 climbed with age in cerebrospinal fluid, with CHIT1 showing a particularly large effect size of 0.450, while plasma levels changed little. Absolute TREM2 concentrations in macaque cerebrospinal fluid were lower than in human samples, hinting at species differences in microglial biology or in how the human-optimized antibodies recognize the macaque protein.

Among synaptic proteins, the neuronal pentraxins NPTX1, NPTX2, and NPTXR declined consistently in the cerebrospinal fluid of older animals, a pattern that echoes findings from human aging and Alzheimer’s cohorts. Crucially, markers of acute neuronal injury such as ENO2, UCHL1, FABP3, and S100B did not rise in parallel, arguing that the pentraxin drop reflects synaptic remodeling rather than wholesale neuronal breakdown. The study also uncovered a possible upstream driver: cerebrospinal pleiotrophin, a growth factor involved in neurite outgrowth and synaptic stabilization, fell markedly with age and remained correlated with all three pentraxins even after adjusting for chronological age. The authors are careful to note that this PTN-pentraxin axis is a correlation, not a demonstrated cause, but it offers a biologically coherent framework for future mechanistic work.

Perhaps the most intriguing result concerns the blood-brain barrier and choroid plexus. Folate receptor 1, or FOLR1, which shuttles folate across the blood-cerebrospinal fluid barrier, and IGFBP7, a cellular senescence marker enriched in the choroid plexus, showed non-linear age patterns with striking inter-individual variance among geriatric animals. Unsupervised clustering revealed that this variance corresponds to two proteomic states: geriatric animals with relatively preserved FOLR1 and IGFBP7 levels, and those with depleted levels. The two states differed across broader biological modules as well, with the high-FOLR1/high-IGFBP7 group showing higher pentraxin and amyloid- and tau-related signals, and the low group showing enrichment of vascular and immune proteins. In plasma, the same clustering analysis produced only weak, unstable partitions, suggesting that structured molecular subtypes, reminiscent of the ageotypes described in human longitudinal studies, emerge specifically within the aging central nervous system.

The neurodegeneration module delivered the study’s starkest divergence from human patterns. Despite confirmed amyloid pathology in the aging brains, cerebrospinal levels of amyloid-beta 38, 40, and 42, brain-derived total tau, and phosphorylated tau species were essentially age invariant, and macaque Aβ42 never reached the depleted levels characteristic of human Alzheimer’s disease. Plasma told a different story: amyloid-beta species and the phosphorylated tau epitopes pTau-181 and pTau-217 rose with age, and plasma amyloid ratios shifted directionally as in human aging. Immunohistochemistry on 41 brains confirmed extensive parenchymal and vascular amyloid deposition in geriatric animals alongside only sparse tau pathology, far milder than in human Alzheimer’s cases. Notably, plasma correlations between CHI3L1, CHIT1, CXCL10, and brain amyloid burden survived correction for chronological age, whereas most cerebrospinal correlations did not, positioning plasma as the more informative window on central amyloid accumulation in this model.

Inflammaging, the chronic low-grade inflammation of aging, was also broadly concordant with humans. Plasma interleukin-6, IL18, IL16, IL33, and the chemokine CCL2 all increased with age, and plasma GDF15, an integrative marker of mitochondrial and cellular stress, showed one of the strongest age associations in the entire panel, with a Spearman correlation of 0.547. Cerebrospinal inflammatory changes were more selective, with IL18 rising strongly and other cytokines elevated only in subsets of geriatric animals. One notable exception emerged: C-reactive protein, which typically rises with age in human cohorts, actually declined with age in macaque plasma despite elevated interleukin-6, a species difference the authors flag for further study. Proteins of metabolic and proteostatic stress, including PARK7, PGK1, and SQSTM1, added further compartment-specific texture to the aging signature.

Inter-individual variability emerged as a unifying theme. Using standardized pairwise distances in high-dimensional biomarker space, the researchers showed that older macaques differed from one another far more than younger ones did, in both fluids, and a sliding-window analysis confirmed a significant positive association between chronological age and molecular dispersion. Yet the rise in dispersion was steeper and more linear in plasma, while cerebrospinal fluid variability increased more gradually and resolved into the partially reproducible subtypes described above. The authors suggest that peripheral aging in macaques is a distributed, multi-pathway process, whereas molecular aging in the brain may follow structured, divergent paths that are not strictly tied to years lived.

The study has honest limitations. The antibody panel was designed for human proteins, and although rhesus macaques share roughly 93 percent genomic sequence identity with humans, some immune-related targets went undetected in macaque fluids, reflecting either true biological divergence or limited antibody cross-reactivity. The cohort is cross-sectional, so temporal ordering cannot be established, and the naturalistic colony includes age-correlated declines in body condition and comorbidities that are themselves part of the aging phenotype. Even so, the work delivers what the field has lacked: a compartment-resolved, human-benchmarked proteomic atlas of primate brain aging. By showing that macaques faithfully reproduce human neurosenescence markers while remaining largely free of terminal neurodegeneration, the study positions these animals as an ideal platform for studying the permissive early phase of brain aging, and for de-risking biomarker-driven clinical trials before they reach patients.

Subject of Research: Proteomic biomarkers of brain aging and early neuropathology in rhesus macaques

Article Title: Plasma and cerebrospinal fluid proteomics reveal distinct patterns of senescence and neuropathology in aging rhesus macaques

Article References: Plasma and cerebrospinal fluid proteomics reveal distinct patterns of senescence and neuropathology in aging rhesus macaques. (n.d.). https://doi.org/10.1007/s11357-026-02555-7

Image Credits: AI Generated

DOI: 10.1007/s11357-026-02555-7

Keywords: rhesus macaque, proteomics, cerebrospinal fluid, plasma biomarkers, neurosenescence, inflammaging, Alzheimer's disease, amyloid-beta, tau, NULISAseq, blood-brain barrier, GeroScience