Age‐equivalent induced neurons from patients reveal neuronal fate loss and metabolic transformation underlying neurodegeneration in sporadic Alzheimer’s disease

Abstract Background Old age is the dominating risk factor for Alzheimer’s Disease (AD), which exclusively affects people at older ages. Sporadic AD represents the overwhelming majority of all cases, but most research on AD has been performed on genetic causes and their directly related pathways, also because we were in lack of models that can reflect complex human genetics and age in a neuronal context. Method Patient‐specific iPSC models represent an attractive solution, but iPSC reprogramming results in cellular rejuvenation and thus yields phenotypically young neurons mirroring fetal development. By contrast, direct conversion of old patient fibroblasts into induced neurons (iNs) preserves endogenous signatures of aging and produces neurons that resemble the adult human brain on an epigenome‐ and transcriptome‐wide scale. To control for the involvement of aging in AD, we generated age‐equivalent fibroblast‐derived iNs and rejuvenated iPSC‐derived neurons from a cohort of AD patients and controls. Result Patient‐derived AD iNs reflect a hypo‐mature neuronal identity characterized by markers of stress, cell cycle, glycolytic reprogramming, and de‐differentiation, which share similarities with malignant cancer transformation and age‐dependent epigenetic erosion. We identified an isoform switch of pyruvate kinase M (PKM) to the PKM2 isoform as a trigger of a Warburg effect‐like metabolic switch, and further as a key transcriptional regulator leading to cell fate instability. Metabolomics, glucose tracing, and acetylome analysis confirmed a hyper‐acetylated state of AD iNs. This affects several splicing factors, triggering an isoform switch to PKM that ultimately leads to a loss of neuronal fate and maturity. Inhibition of the resulting citrate shunt to acetylCoA modulates the isoform switch in favor of PKM1, and re‐instates a mature and more resilient neuronal state. Conclusion We conclude that age‐equivalent adult‐like iNs from patients are a useful model to study age‐dependent pathological signatures of AD. This iN model identifies AD‐related neuronal changes as part of an active cellular program that impairs neuronal cell identity and neuronal resilience by utilizing cancer‐related signaling cascades. We further reason that the AD‐specific metabolic switch and fate loss in neurons regulates, and is reinforced by, alternative splicing of PKM.