Back

An age-preserving neuronal iSpheroid platform for modeling human aging and Alzheimer disease

Alsolami, S.; Wang, M.; Sarraf Sabedot, T.; Singh, U.; Kirk, B.; Mitchell, L.; Hernandez, M.; Skvir, N.; Bowman, C.; Oiwa, K.; Agarwal, R.; Mahmoudi, M.; Cuoco, M. S.; MacDonald, E.; Cranston, S.; Lucciola, R.; Boassa, D.; Herdy, J. R.; Jones, J. R.; Yoshikiyo, K.; Mertens, J.; Martinez-Trujillo, J.; Jaremko, L.; Murphy, G. J.; Gage, F. H.

2026-10-01 neuroscience
10.64898/2026.09.30.755853 bioRxiv
Show abstract

Aging is the strongest risk factor for Alzheimer disease (AD), yet existing human neuronal models either erase age-associated features during reprogramming or fail to recapitulate three-dimensional organization of human neuronal tissue. Here, we establish directly induced neuronal spheroids (iSpheroids), a scalable three-dimensional platform generated by direct conversion of primary fibroblasts without pluripotency. iSpheroids generate long-lived, electrophysiologically active neurons from young, unimpaired aged, and sporadic AD donors while preserving donor-specific epigenetic age. Single-nucleus RNA sequencing revealed approximately 96% of cells acquired a neuronal identity and were distributed across multiple neuronal states, including a diversity of neurons. The platform supports incorporation of astrocytes and microglia and generation from non-human primate fibroblasts. AD iSpheroids spontaneously recapitulated neuronal degeneration, amyloid-beta; accumulation, progressive Tau pathology, and metabolic dysfunction. Integrated proteomic, metabolomic, and lipidomic profiling distinguished physiological aging from AD and revealed disease-associated metabolic remodeling. Oxaliplatin shifted AD iSpheroids toward oxidative metabolism, whereas a multi-compound screen identified trametinib as a modulator of inflammatory and age-associated programs.

Matching journals

The top 6 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.