Spatial Omics in Neurodegenerative Drug & Biomarker Discovery

Jonathan Alles

EVOBYTE Digital Biology

By EVOBYTE Your partner in bioinformatics

Introduction

Drug discovery for neurodegenerative diseases often feels like trying to diagnose a citywide power outage from a single streetlight. Traditional bulk omics tell us which genes are on or off, but not where and alongside whom. Spatial omics changes that. By measuring RNA, proteins, and even metabolites with their native histological coordinates, spatial omics reveals the cellular neighborhoods where pathology begins, spreads, and stalls.

This installment of our Bioinformatics Strategy Series focuses on spatial omics for neurodegenerative disease research. We’ll look at how spatial transcriptomics (ST), spatial proteomics, and spatial multi‑omics uncover biomarker candidates and therapeutic targets in Alzheimer’s disease (AD), Parkinson’s disease (PD), and amyotrophic lateral sclerosis (ALS). Along the way, we’ll acknowledge a practical truth: brain specimens are precious, heterogeneous, and limited. Yet with the right analytic playbook, even small cohorts can yield actionable biology.

Why spatial omics, and why now?

The brain is exquisitely organized. Layers in cortex, nuclei in the midbrain, axon tracts that run like highways—every feature is a clue. Neurodegeneration doesn’t strike randomly; it maps onto these features. Spatial omics preserves that map.

Spatial transcriptomics (ST) captures gene expression in situ, typically on platforms like 10x Visium or NanoString GeoMx. Spatial proteomics extends the idea to proteins via multiplexed immunofluorescence or imaging mass spectrometry. When you integrate these modalities with single‑nucleus RNA‑seq (snRNA‑seq), histopathology, and neuropath staging, a picture emerges: discrete cellular states cluster around plaques, tangles, Lewy pathology, or dying motor neurons. Those states, and the molecules they express, are where drug targets and biomarkers hide.

In practice, this means moving from “what is different” to “what is different, where, and in which microenvironment.” That shift is more than academic. Therapies aimed at a cell state that exists only near amyloid plaques will fail if they’re delivered broadly or measured in the wrong tissue region. Spatial data de‑risks that mismatch early.

From signals to sites: what spatial omics is revealing in Alzheimer’s disease

Alzheimer’s disease is the proving ground for spatial approaches. Multiple human‑tissue studies now show that glial cells adopt distinct, location‑specific phenotypes within plaque niches. Disease‑associated microglia (DAM) and disease‑associated astrocytes (DAA) coalesce around plaques, turning up genes tied to lipid handling, phagocytosis, complement activation, and cytokine signaling. This localized immune remodeling tracks with synaptic loss, neuronal stress markers, and gradients of plaque morphologies, suggesting a cause‑and‑consequence dance in the immediate neighborhood of pathology. These spatial signatures point to targetable pathways—complement, microglial lipid metabolism, and interferon signaling—that could be modulated with greater precision when we know where they’re active.

Layer‑aware maps add another layer of insight. Using spatial transcriptomics in the human middle temporal gyrus, researchers have traced vulnerability to AD across cortical laminae, highlighting hub genes and pathways that differ by layer and white matter adjacency. These layer‑specific signals, validated down to single molecules, explain why seemingly similar tissue punches can tell very different molecular stories and reinforce the need to sample deliberately when building biomarker assays or evaluating drug response.

Spatial omics is also surfacing interactions between microglia and astrocytes that may convert local immune responses from protective to toxic. In an amyloid mouse model, astrocytes in microglia‑dense plaque niches acquired a more neurotoxic signature and disrupted neuronal signaling. This microenvironmental crosstalk—visible only when you keep the tissue architecture intact—offers a rationale for combination strategies that rebalance both glial partners rather than targeting one cell type in isolation.

Taken together, these findings don’t just nominate genes; they nominate neighborhoods. That’s crucial for companion diagnostics. If a plasma biomarker reports on a microglial state that exists primarily within 100 microns of plaques, then clinical readouts and trial enrichment criteria should reflect that biology. Spatial maps help align those decisions.

Beyond Alzheimer’s: spatial insights in Parkinson’s disease and ALS

In Parkinson’s disease, the substantia nigra pars compacta (SNpc) is ground zero, but vulnerability is not uniform. Spatial transcriptomics in human SNpc has started to distinguish molecular patterns across dorsal‑ventral and medial‑lateral axes, teasing apart signatures shared with multiple system atrophy (MSA) and those unique to PD. The patterns implicate protein synthesis, mitochondrial function, immune processes, and the ubiquitin‑proteasome system, and they point to hubs like amyloid precursor protein in network analyses. For target discovery, that granularity matters: a mitochondrial pathway signal centered in a ventrolateral quadrant suggests different delivery strategies and off‑target risks than a diffuse brain‑wide signal.

Cell‑type context adds another layer. Single‑nucleus atlases of the human SN reveal neuron classes and glia with distinct PD risk enrichments, reminding us that not every transcript change in nigra is dopaminergic. When spatial analyses show where these risk‑laden cell types are concentrated, teams can design experiments that prioritize those regions for validation and model building.

ALS presents a different challenge: upper and lower motor neuron degeneration unfolds across cortical layers and spinal cord tracts. Spatial enrichment studies in human motor cortex link ALS genes to layer 5 excitatory neurons and nominate candidates such as NOMO1 through rare‑variant burden analyses coupled to spatial expression. That design—genetics meets spatial biology—illustrates a powerful playbook for discovering biomarkers like neurofilament drivers, then tracing their spatial origins to the neurons and circuits that fail first.

What unites these disease areas is a strategic shift from cell counts to microenvironments. By prioritizing where distinct cell states congregate—and how close they are to hallmark pathology—we extract mechanistic clues that bulk omics averages away.

Real‑world examples that point to biomarkers and targets

Alzheimer’s plaque‑glia niches have revealed consistent molecular themes. Spatial studies in human cortex show DAM and DAA concentrated around plaques, with upregulated complement and lipid handling programs alongside synaptic vulnerability markers. In vitro models using iPSC‑derived microglia mirror these plaque‑adjacent glial states, lending translational credibility and offering a tractable test bed for pathway perturbation. Here, C1q‑driven complement and lipid metabolism genes in microglia rise as plausible therapeutic levers and as anchors for proximity‑aware biomarker panels.

Layer‑resolved maps in the middle temporal gyrus add specificity. Novel layer‑linked differentially expressed genes—validated by single‑molecule FISH—highlight why a biomarker measured from temporal cortex might outperform one from frontal cortex for early detection, and why regional sampling protocols matter in clinical studies. This level of detail supports more precise patient selection and endpoint design, especially when layering on plasma markers that trace back to specific cortical layers.

Microglia–astrocyte crosstalk around plaques underscores the case for combination therapy. Spatial data show astrocytes acquiring neurotoxic features in microglia‑dense plaque niches, which correlates with disrupted neuronal signaling. If a drug shifts microglia toward a reparative state but leaves astrocytes unchecked, the net benefit may be muted. Designing trials to read out both microglial and astrocytic state changes—ideally with spatially anchored tissue endpoints—can reduce that risk.

In Parkinson’s disease, spatial transcriptomics of the SNpc captures quadrant‑specific signatures involving mitochondrial function, protein turnover, and immune processes, with network analyses nominating targets that connect these pathways. When these maps are integrated with single‑nucleus atlases that assign PD genetic risk to specific nigral cell types, you get a short list of hypotheses with anatomical coordinates. That’s a powerful starting point for precision delivery approaches or for imaging biomarkers tuned to vulnerable nigral subregions.

ALS benefits from a genetics‑meets‑spatial approach. Studies combining spatial expression in motor cortex with rare‑variant analyses have nominated NOMO1 and reinforced the centrality of NEFL as a marker tied to layer 5 motor neuron biology. This recipe—map the spatial origin of genetic risk and biomarker signals—helps teams design assays that are mechanistically grounded and prioritize targets with a clear site‑of‑action narrative.

Summary / Takeaways

Spatial omics brings location into focus, and location is often the missing variable in neurodegeneration. For biomarker discovery, it clarifies which signals truly originate from disease‑critical microenvironments and which are background noise. For target discovery, it anchors pathway hypotheses to the neighborhoods where drugs must act. And for translational strategy, it informs where and how to sample scarce tissue to maximize discovery power.

If you’re planning your next study, start by asking three questions. What microenvironment do we believe drives this disease feature? Which modality—transcript, protein, or both—will best capture the biology in that niche? And how will we connect spatial endpoints to peripheral biomarkers or imaging readouts in the clinic? Teams that answer these early, and build an analysis loop that respects specimen limitations, move faster with fewer surprises.

What brain neighborhood do you want to map next—and what decision will that map change?