A research team led by Qianqian Li, Jiale Chen, and Miao Ren of the Hainan University collaboration with the HUST-Suzhou Institute for Brainsmatics dissected specific subregions of the mouse basal forebrain, midbrain, and brainstem in both adult animals aged three to five months and aged animals aged eighteen to twenty-four months. Using the 10x Genomics Chromium platform, they generated single-nucleus RNA sequencing libraries and, after rigorous quality control, assembled an atlas of 108,062 high-quality cells. Computational annotation based on canonical marker genes identified eleven major cell types, including GABAergic and glutamatergic neurons, cholinergic neurons, oligodendrocytes and their precursors, microglia, astrocytes, endothelial cells, ependymal cells, pericytes, and fibroblasts. Reference-based label transfer with Seurat confirmed the assignments with high concordance, giving the team a reliable cellular census across three anatomically and functionally distinct cholinergic territories.

The first striking finding came from comparing how entire brain regions respond to age at the multicellular level. In the aged basal forebrain, upregulated differentially expressed genes were consistently enriched across multiple cell types in pathways tied to p53-mediated inflammatory responses, oxidative phosphorylation, synaptic translation, and protein synthesis. No comparable coordinated transcriptional activation appeared in the aged midbrain or brainstem. Meanwhile, genes downregulated with age in all three regions clustered around axon guidance and neuronal maintenance. In other words, the basal forebrain undergoes a robust, brain-wide transcriptional remodeling characterized by heightened stress and metabolic programs alongside erosion of the pathways that keep neurons structurally intact, a pattern conspicuously absent from its more resilient neighbors.

To probe how cells talk to one another, the researchers applied CellChat, a computational framework that infers intercellular communication from ligand-receptor expression. The analysis revealed a striking regional specificity. The ApoE signaling pathway, a major genetic risk factor for Alzheimer’s disease, was active exclusively in the basal forebrain, where microglia acted as the principal receivers of ApoE signals emitted by cholinergic neurons and other cell types. The aged brainstem instead showed specific activation of insulin-like growth factor signaling, a change the authors interpret as a potentially protective or compensatory mechanism supporting metabolic homeostasis and survival. The aged midbrain displayed enhanced protocadherin signaling, associated with neuronal identity and connectivity. These distinct signaling landscapes suggest that each cholinergic system ages within its own extracellular milieu, shaped by different glial and metabolic partners.

One pathway, however, spanned all three regions: amyloid precursor protein, or APP, signaling, with cholinergic neurons identified as the putative receiver cells. Here the regional differences proved decisive. In the basal forebrain and midbrain, the importance of APP signaling to cholinergic neurons rose substantially with age, driven largely by the App-Tnfrsf21 ligand-receptor pair, whereas the brainstem showed no such age-related increase. Given that APP processing lies at the heart of amyloid-beta generation in Alzheimer’s disease, the selective amplification of APP reception in aged basal forebrain cholinergic neurons hints at an extracellular signaling context that may prime these cells for degeneration. The finding dovetails with recent work showing that aging selectively increases the vulnerability of the cholinergic system to amyloid pathology.

Zooming in on the cholinergic neurons themselves, unbiased subclustering revealed that they segregate cleanly by anatomical origin, forming the most regionally distinctive transcriptional profile of any cell type in the atlas. Basal forebrain cholinergic neurons highly expressed Lhx8, Ngfr, and Ntrk1, the latter two encoding receptors for nerve growth factor, a trophic signal long known to sustain these cells. Midbrain neurons were enriched for Tgfbr2, Egflam, and Nos1, while brainstem neurons expressed Sv2b, Ryr3, and Nfia. RNAscope in situ hybridization validated the star of this list: Ngfr was expressed in nearly all basal forebrain cholinergic neurons and virtually absent from the midbrain and brainstem populations. Functional enrichment underscored the divide, with basal forebrain genes tied to neurodegenerative disease pathways and synaptic signaling, midbrain genes to nitric oxide and Wnt signaling, and brainstem genes to longevity-regulating and PI3K-Akt pathways.

High-dimensional weighted gene co-expression network analysis then exposed how aging rewrites these programs differently in each region. In the basal forebrain, an age-correlated module was enriched in neurodegenerative disease pathways, oxidative phosphorylation, ATP biosynthesis, and apoptotic signaling, and contained the aging-linked genes Malat1 and Ube3a. The midbrain’s age module pointed instead toward RNA processing, nuclear transport, and chromatin remodeling, while the brainstem’s declining module involved neuronal migration, axon guidance, and neurotransmitter transport. Gene set variation analysis confirmed the asymmetry: stress and apoptosis pathways climbed in aged basal forebrain cholinergic neurons, whereas metabolic homeostasis and growth signaling pathways, including insulin receptor recycling and IGF1 signaling, rose in the midbrain and brainstem while falling in the basal forebrain.

Differential expression between adult and aged cholinergic neurons within each region revealed remarkably little overlap, with 186 genes uniquely upregulated and 92 uniquely downregulated in the aged basal forebrain, and hundreds of region-specific changes in the brainstem but none in the midbrain. The basal forebrain losses were especially telling, encompassing axonal and cytosolic transport, neurotrophic signaling, lipid membrane metabolism, and the somatic and synaptic architecture that keeps a neuron connected. The brainstem, by contrast, upregulated gene sets for ion homeostasis, energy metabolism, cytoskeletal maintenance, and cell survival, an adaptive profile consistent with resilience. Two genes with opposite regional trajectories stood out. Oxr1, a conserved antioxidant protein that protects mitochondria and guards against oxidative damage, rose with age in basal forebrain cholinergic neurons and fell in the brainstem. Ptprs, a receptor tyrosine phosphatase governing axon growth, adhesion, and synapse formation, dropped in the basal forebrain and trended upward in the brainstem.

Using the machine-learning framework scTenifoldKnk, the team simulated virtual knockouts of these two candidates in basal forebrain cholinergic neurons. Perturbing Oxr1 disrupted 81 downstream genes, including 22 that were both age-related and hub genes of the oxidative stress module, with enrichment in the ERK1/2 cascade, synapse assembly, and neuron projection extension. Virtual knockout of Ptprs perturbed 69 genes, among them neurodevelopmental regulators such as Nfib and Reln, converging on axon guidance, synaptic organization, and neurotransmitter cycling. Both perturbations thus struck the same target: the gene networks that maintain neuronal structure and synaptic function. The authors are careful to note the caveats. Tissues from multiple mice were pooled per library, limiting replicate-level statistics; both sexes were combined, precluding assessment of sex-specific effects; and the findings remain transcriptomic predictions awaiting experimental validation in living models. Many brainstem cholinergic neurons are also motor neurons with fundamentally different physiological demands, which may partly explain their apparent resilience. Even so, the study delivers a compelling molecular framework for one of neuroscience’s enduring questions: why the neurons that sustain our memory and attention are precisely the ones that time attacks first, and where, in their shifting gene networks, the seeds of that vulnerability lie.

Subject of Research: Region-specific molecular mechanisms underlying selective aging vulnerability of cholinergic neurons in the mouse basal forebrain, midbrain, and brainstem

Article Title: Single-nucleus transcriptomics uncovers region-specific molecular features underlying selective aging vulnerability of cholinergic neurons

Article References: Li, Q., Chen, J., Liu, X., Gong, H., Li, X., & Ren, M. (2026). Single-nucleus transcriptomics uncovers region-specific molecular features underlying selective aging vulnerability of cholinergic neurons. iScience, 29(11), Article 117778. https://doi.org/10.1016/j.isci.2026.117778

Image Credits: AI Generated

DOI: 10.1016/j.isci.2026.117778

Keywords: cholinergic neurons, basal forebrain, single-nucleus RNA sequencing, brain aging, Alzheimer's disease, ApoE signaling, APP signaling, Oxr1, Ptprs, neurodegeneration, CellChat, transcriptomics

The research team, led by Sungwoo Kang and Byoung Seok Ye of Yonsei University Severance Hospital in Seoul, analyzed 387 patients with amyloid-confirmed Alzheimer’s disease spanning the full clinical spectrum from mild cognitive impairment to dementia. All participants underwent detailed neuropsychological testing, 3.0-Tesla structural MRI, amyloid PET with florbetaben, and dual-phase FP-CIT PET imaging. The cohort was divided into 205 APOE4 carriers and 182 non-carriers, and the researchers carefully excluded 58 patients who showed two or more core clinical features of Lewy body disease, such as parkinsonism, cognitive fluctuation, REM sleep behavior disorder, or recurrent visual hallucinations, to ensure the cohort remained as Alzheimer’s-consistent as possible.

Two imaging biomarkers anchored the analysis. The first was basal forebrain volume, measured from the Ch4 region corresponding to the nucleus basalis of Meynert, the principal source of acetylcholine to the cortex and a structure long known to degenerate in Alzheimer’s disease. The second was striatal dopamine transporter uptake, quantified from late-phase FP-CIT PET scans, with the posterior caudate selected as the representative dopaminergic marker because it showed the strongest relationship with basal forebrain volume. Regional brain perfusion was derived from early-phase FP-CIT PET using a data-driven normalization method called the subject residual profile, which captures each patient’s region-relative perfusion deviations without relying on a potentially disease-affected reference region such as the pons or cerebellum.

The first striking result was a clean genetic dissociation of the two neurotransmitter systems. APOE4 carriers showed significantly lower basal forebrain volume than non-carriers, consistent with decades of autopsy work linking the ε4 allele to reduced cholinergic activity in the nucleus basalis and greater loss of presynaptic cholinergic markers in the neocortex. Non-carriers, by contrast, showed lower posterior caudate dopamine transporter uptake, indicating more pronounced nigrostriatal dopaminergic degeneration. Both patient groups had reduced basal forebrain volume and striatal transporter uptake compared with 53 healthy controls, but the relative burden of damage flipped according to genotype, a pattern confirmed by statistically significant interaction terms between APOE4 status and each biomarker.

The perfusion maps told a parallel story. In APOE4 carriers, lower basal forebrain volume was associated with hypoperfusion across the temporoparietal association cortices and the posterior cingulate cortex, the classic Alzheimer’s signature, in addition to medial temporal changes seen in both groups. In non-carriers, lower dopamine transporter uptake was linked to a broader set of perfusion alterations, including frontal hypoperfusion and relative hyperperfusion in the amygdala, hippocampus, and ventral occipitotemporal regions, patterns that echo the metabolic signatures of Lewy body pathology. The researchers suggest that in carriers, Lewy-type alpha-synuclein involvement may be expressed more through cholinergic and cortical-limbic pathways, whereas in non-carriers, brainstem-predominant Lewy pathology may drive the dopaminergic signal, consistent with autopsy studies showing genotype-dependent differences in the anatomical distribution of Lewy-related pathology.

When the team connected these biomarkers to symptoms, the divergence sharpened. In APOE4 carriers, lower basal forebrain volume was associated with poorer memory performance and with more severe neuropsychiatric symptoms including hallucinations, delusions, anxiety, apathy, aberrant motor behavior, and appetite changes. Mediation analyses showed that the link between cholinergic degeneration and memory dysfunction was statistically mediated by posterior cingulate hypoperfusion, meaning the structural damage appeared to act on cognition through its effect on cortical function. After accounting for regional perfusion, direct associations remained for hallucinations, delusions, and apathy, suggesting that cholinergic loss contributes to these behavioral disturbances through mechanisms beyond cortical hypoperfusion alone.

In non-carriers, the picture was almost inverted. Lower posterior caudate dopamine transporter uptake was associated with poorer performance across all cognitive domains measured, including attention, language, visuospatial ability, memory, and executive function, and with more severe hallucinations, delusions, and anxiety. Remarkably, these associations largely persisted even after adjustment for regional perfusion, indicating that dopaminergic degeneration in this group exerts direct effects on cognition and behavior rather than acting purely through cortical metabolic decline. Mediation analyses identified significant indirect pathways through angular gyrus hypoperfusion for memory and executive function, but the direct associations remained significant alongside them. The researchers also observed a synergistic interaction in non-carriers, where the association between dopaminergic loss and delusion severity was strongest in patients who also had the smallest basal forebrain volumes.

The memory findings deserve particular attention because they challenge the assumption that memory impairment in Alzheimer’s always reflects the same circuit damage. In APOE4 carriers, memory dysfunction tracked hypoperfusion in the posterior cingulate cortex and medial temporal lobe, regions densely connected to the cholinergic basal forebrain and heavily burdened by tau in carriers. In non-carriers, memory problems correlated most strongly with parietal hypoperfusion, particularly in the angular gyrus, a region more often implicated in atypical, non-amnestic presentations. This topographic split aligns with prior imaging work showing greater medial temporal atrophy and tau accumulation in carriers versus greater frontoparietal cortical thinning and tau burden in non-carriers, and it helps explain why non-carriers more frequently present with atypical clinical syndromes.

The study’s authors are careful about its limits. The cross-sectional design cannot establish causality, and the possibility that chronic clinical dysfunction drives secondary hypoperfusion and subcortical atrophy cannot be excluded, although prior longitudinal work shows that basal forebrain degeneration precedes and predicts the cortical spread of Alzheimer’s pathology. Basal forebrain volume is also an indirect structural surrogate rather than a direct measure of cholinergic synaptic activity, though multimodal studies have validated it against vesicular acetylcholine transporter and acetylcholinesterase PET. The cohort was enriched for patients referred for FP-CIT PET, potentially inflating dopaminergic abnormalities, and tau and alpha-synuclein pathology were not directly assessed. Effect sizes were modest, and the authors emphasize the need for validation in independent cohorts with molecular biomarkers of tau, synuclein, and TDP-43 pathology.

Even with those caveats, the implications are substantial. If APOE4 genotype defines two neurochemical subtypes of amyloid-confirmed Alzheimer’s disease, then cholinergic and dopaminergic imaging markers could become powerful tools for patient stratification in anti-amyloid trials, where treatment responses have been notoriously heterogeneous. The findings also raise the prospect of genotype-guided symptomatic therapy: cholinesterase inhibitors may matter most for the basal-forebrain-dominant carrier subtype, while dopaminergic pathways may deserve greater attention in non-carriers. As the authors conclude, combining APOE4 status with neurotransmitter imaging may move the field closer to a precision medicine framework in which the neurochemical fingerprint of each patient’s disease, not just its amyloid burden, determines the therapeutic strategy.

Subject of Research: APOE4-dependent cholinergic and dopaminergic degeneration and their relationships to cognitive and neuropsychiatric symptoms across the Alzheimer's disease spectrum

Article Title: APOE4-dependent cholinergic/dopaminergic contributions to clinical symptoms across Alzheimer’s disease spectrum: A retrospective observational study

Article References: Kang, S., Jeon, S., Lee, D., Lee, H., Jeon, S.-H., Choi, M., Lee, Y.-G., Shin, N.-Y., Yun, M., & Ye, B. S. (2026). APOE4-dependent cholinergic/dopaminergic contributions to clinical symptoms across Alzheimer’s disease spectrum: A retrospective observational study. European Journal of Nuclear Medicine and Molecular Imaging. https://doi.org/10.1007/s00259-026-08165-x

Image Credits: AI Generated

DOI: 10.1007/s00259-026-08165-x

Keywords: Alzheimer's disease, APOE4, cholinergic system, dopaminergic system, basal forebrain, dopamine transporter, FP-CIT PET, brain perfusion, neuropsychiatric symptoms, memory dysfunction, Lewy body pathology, biomarkers