The nose tells before memory: loss of smell in Alzheimer's begins with the breakdown of norepinephrine fibers

Olfaction is one of the most sensitive indicators of the onset of Alzheimer's disease. A new paper in Nature Communications shows that the key to early loss of smell is not in the cortex or amyloid plaques, but in the very "entrance" of the olfactory system: mice with amyloid pathology lose some of the norepinephrine axons from the locus coeruleus (LC) in the olfactory bulb long before plaques appear, and this is what disrupts the perception of smells. The mechanism is unpleasantly simple: microglia recognize a "disposal mark" on these axons and phagocytose them. Genetic weakening of this "eating" preserves the axons - and the sense of smell. In people with a prodromal stage, the authors find a similar picture according to the PET biomarker of microglia and postmortem histology.

Background

Early loss of smell is one of the most consistent harbingers of neurodegeneration. It is well known from Parkinson's disease, but in Alzheimer's disease (AD), hyposmia often appears before noticeable memory lapses. Until now, the main focus of explanations has been "cortical-amyloid": it was believed that the deterioration of smell is a side effect of Aβ/tau accumulation and cortical dysfunctions. However, the olfactory system does not originate in the cortex, but in the olfactory bulb (OB), and its work is fine-tuned by ascending modulatory systems, primarily the noradrenergic projection from the locus coeruleus (LC).

LC is the first "node" of the brain involved in AD: according to postmortem data and neuroimaging, its vulnerability is recorded already at the prodromal stages. Norepinephrine from LC increases the signal-to-noise ratio and "learning" plasticity in the OB; this means that the loss of the LC input can directly spoil the encoding of odors even before cortical changes. In parallel, microglia, the brain's immune cells, are on the scene. Normally, they "trim" synapses and remove damaged network elements, recognizing "disposal marks" on membranes (for example, external phosphatidylserine). In chronic stress and protein failures, such "sanitation" can turn into excessive phagocytosis, depriving the network of working conductors.

Taken together, this forms an alternative hypothesis for early hyposmia in AD: not plaques per se, but a selective vulnerability of the LC→OB pathway plus microglial axonal 'cleaning'. This idea is biologically sound, but until recently there was a lack of direct evidence on key points:

1. does the decay begin with the LC axons (and not with the death of the LC neurons themselves),
2. does this happen very early and locally in the OB,
3. does microglial phagocytosis play a leading role, and
4. whether human correlates are visible - from olfactory tests, PET microglia markers and histology.

Hence, the goals of the current study are to disentangle structural wiring loss from LC “weak activation,” to disentangle the contributions of amyloid and immune clearance, to demonstrate causality using genetic inhibition of phagocytosis, and to correlate the mouse findings with early AD in humans. If the “weak link” does indeed lie along the LC→OB pathway, this opens up three practical directions: prodrome network biomarkers (simple olfactory tests + targeted bulbar neuroimaging), new intervention points (modulation of microglia’s “eat-me” signal recognition), and a paradigm shift in early diagnosis from “ubiquitous amyloid” to the vulnerability of specific neural networks.

What exactly did they find?

• The earliest hit is to the olfactory bulb. In the App NL-GF model, the first signs of LC axon loss appear between 1-2 months and reach ~33% fiber density loss by 6 months; in the hippocampus and cortex, decay begins later (after 6-12 months). At this stage, the number of LC neurons themselves does not change - it is the axons that suffer.
• Not "all modalities in general", but selectively LC→OB. Cholinergic and serotonergic projections in the olfactory bulb do not thin out in the early stages, which indicates the specificity of the lesion of the norepinephrine system.
• Behavior confirms the mechanism. Mice are less successful at finding hidden food and less willing to explore a scent (vanilla) by 3 months - the earliest behavioral manifestation described in this model.
• Not a basal NA, but a "phase response". Using the fluorescent sensor GRAB_{NE}, it was shown that the odor of sick mice causes an evoked release of norepinephrine in the bulb for different odorants.
• Microglia "eat" LC axons. The key trigger is external exposure of phosphatidylserine on axon membranes; microglia recognize this "tag" and phagocytose the fibers. Genetic reduction of phagocytosis preserves LC axons and partially preserves olfaction.

An important detail: the early loss of LC fibers in the olfactory bulb is not associated with the amount of extracellular Aβ at the same time. This shifts the focus from the “plaques” to the vulnerability of the specific network and immune cleanup. And an attempt to “turn up the volume” of the remaining LC axons chemogenetically did not restore the behavior - so it is not just a matter of weak activation, but a structural loss of wiring.

What was shown in people

• PET signature of microglia in the olfactory region. Patients with prodromal Alzheimer's disease (SCD/MCI) have increased TSPO-PET signal in the olfactory bulb - similar to early diseased mice. This, judging by the mouse/human comparison, reflects a higher density of microglia, and not just their "activation".
• Histology confirms the loss of LC fibers. In postmortem samples of the olfactory bulb, early Alzheimer's cases (Braak I-II) have lower NET+ (LC axon marker) density than healthy peers. In later stages, it does not decrease further - the early "window of vulnerability" has already closed.
• Olfactory tests "mature" together with the process. In the prodrome, a tendency toward hyposmia is visible, with a manifest diagnosis - a reliable deterioration in odor identification.

Why is this important?

• Early diagnostic window: Combining simple olfactory tests with targeted neuroimaging (e.g. TSPO-PET of the olfactory bulb) can detect network-specific changes before cognitive complaints occur.
• A new application point for therapy. If hyposmia in Alzheimer is triggered by microglial phagocytosis of LC axons, then the targets are the signaling pathways for recognizing phosphatidylserine and “eating” axons. Stopping this process at early stages means potentially preserving network function.
• Paradigm shift. Not all early symptoms are dictated by amyloid: vulnerability of specific neural networks (LC→OB) and “sanitary” processes of the immune system may be more primary in time.

A little physiology to connect the dots

• The locus coeruleus is the main source of norepinephrine for the forebrain; it regulates wakefulness, attention, memory, and sensory filtering, including olfaction. Its integrity is an early predictor of cognitive decline.
• The olfactory bulb is the first odor "comparator"; norepinephrine from the LC fine-tunes its work, including odor learning. Loss of input → worse signal-to-noise ratio → hyposmia.
• Microglia are the brain's "immune gardeners": normally they trim synapses and remove debris. But if phosphatidylserine (usually hidden inside the membrane) appears on an axon, it's like a "dispose of" label - and the network branch is lost.

What does this mean in practice - today

• Consider olfactory screening in people at risk (family history, complaints of "missing smells") and in mild cognitive impairment - it is cheap and informative.
• Research protocols should include olfactory testing and TSPO-PET of the olfactory bulb as early markers of network vulnerability.
• Early stage pharmacology must look not only at amyloid/tau, but also at the LC↔microglia↔olfactory bulb axis - from phosphatidylserine recognition receptors to regulators of phagocytosis.

Restrictions

• Mouse ≠ human. The underlying mechanics are shown in the model; humans have supporting evidence (TSPO-PET, postmortem sections), but the causal chain needs to be proven in clinical studies.
• Small human cohorts. TSPO-PET was performed in a small group; the relationship of bulbar signal level to olfactory dynamics remains to be clarified.
• The difficulty of targeting microglia. It is impossible to completely "switch off" phagocytosis - the brain needs it. The question is in fine tuning and the correct phase of the disease.

Conclusion

In Alzheimer's, "missing smells" may be a direct consequence of early loss of LC norepinephrine fibers in the olfactory bulb, driven by microglia; this opens the door to network biomarkers and early intervention before significant memory loss occurs.

Source: Meyer C. et al. Early Locus Coeruleus noradrenergic axon loss drives olfactory dysfunction in Alzheimer's disease. Nature Communications, 8 August 2025. Open access. https://doi.org/10.1038/s41467-025-62500-8