How is spatial biology used in neuroscience drug discovery?

Spatial transcriptomics and related methods measure gene expression while preserving each cell’s precise anatomical location within the brain, revealing which genes are active in specific vulnerable regions and how molecular signals change across brain areas over time. This has identified region-specific disease signatures, including circadian disruption preceding Alzheimer’s pathology and gene modules linked to regional vulnerability in Parkinson’s disease.

Can spatial transcriptomics map brain disease?

Yes. Spatial transcriptomics has mapped region-specific circadian rhythm disruption in an Alzheimer’s disease mouse model that occurs before substantial amyloid plaque deposition, and has identified specific genes and functional gene modules associated with regional vulnerability to Parkinson’s disease using human brain atlas data integrated with post-mortem tissue.

What are region-specific CNS targets?

Genes or pathways whose relevance to a neurological disease depends specifically on which brain region is being examined, since disease vulnerability in conditions such as Parkinson’s and Alzheimer’s disease concentrates in specific anatomical regions rather than affecting the whole brain uniformly. Identifying these targets requires spatially resolved data, since dissociated sequencing discards the regional origin of each measurement.

Spatial biology in CNS drug discovery

Highlights in this article

Why CNS needs spatial approaches
Mapping neurodegeneration
Region-specific targets
The neural microenvironment
Model systems and translation

Improving CNS drug discovery outcomes means confronting a specific, well-documented problem: this therapeutic area fails more often, and takes longer to succeed, than almost any other in pharma, and a meaningful part of that problem traces back to studying brain disease with methods that discard the one property, precise anatomical position, that the brain relies on most for normal function.


Key takeaways CNS drugs reach approval at less than half the rate of non-CNS drugs, 6.2% versus 13.3% in one Tufts analysis, and take roughly 18% longer to develop. A separate analysis estimated approximately 99.6% attrition for Alzheimer’s disease drug programs specifically between 2002 and 2012. Spatial transcriptomics has identified region-specific circadian rhythm disruption in an Alzheimer’s mouse model that precedes visible amyloid plaque deposition, a potential early signal rather than one requiring established pathology. A separate study named specific genes and two functional gene modules associated with regional vulnerability in Parkinson’s disease, both concentrated in brain regions involved in preclinical disease stages. Poor translation from rodent models to human CNS biology remains a persistent, specific barrier that spatial data is only beginning to address.

Key takeaways

CNS drugs reach approval at less than half the rate of non-CNS drugs, 6.2% versus 13.3% in one Tufts analysis, and take roughly 18% longer to develop.
A separate analysis estimated approximately 99.6% attrition for Alzheimer’s disease drug programs specifically between 2002 and 2012.
Spatial transcriptomics has identified region-specific circadian rhythm disruption in an Alzheimer’s mouse model that precedes visible amyloid plaque deposition, a potential early signal rather than one requiring established pathology.
A separate study named specific genes and two functional gene modules associated with regional vulnerability in Parkinson’s disease, both concentrated in brain regions involved in preclinical disease stages.
Poor translation from rodent models to human CNS biology remains a persistent, specific barrier that spatial data is only beginning to address.

Why CNS needs spatial approaches

The attrition numbers for CNS drug discovery are worth stating precisely, since a vague sense that "CNS is hard" understates how specific and severe the problem actually is. A Tufts Center for the Study of Drug Development analysis found the overall clinical approval success rate for CNS compounds first tested in human subjects from 1995 to 2007 was 6.2%, less than half the 13.3% rate for non-CNS drugs over the same period, and CNS drugs approved for marketing between 1999 and 2013 took approximately 18% longer to reach approval than non-CNS compounds. A separate, more recent review states that a ClinicalTrials.gov analysis of Alzheimer’s disease programs specifically, covering 2002 to 2012, estimated approximately 99.6% attrition, a disease-specific figure even more severe than the general CNS rate.

A substantial share of that failure traces to a problem spatial biology is specifically built to address: brain function depends on precise anatomical position in a way few other organ systems do, since a neuron’s identity and behavior are shaped by exactly which circuit and which anatomical region it belongs to, information dissociation for bulk or standard single-cell sequencing destroys completely. A finding true of neurons from one brain region can be entirely misleading if silently generalized to a different region with a different local circuit and different disease vulnerability.

Mapping neurodegeneration

The clearest demonstration that spatial resolution reveals disease-relevant signal invisible to non-spatial methods comes from a recent finding about the timing of Alzheimer’s pathology relative to a specific, measurable brain function.


A disease signature that shows up before the pathology does A recent study used large-scale spatial transcriptomics to map 24-hour rhythmic gene transcription across cortical and subcortical regions of the mouse brain, finding marked regional differences in rhythmicity, including distinct oscillatory signatures across cortical areas and along the front-to-back axis of the brain. In the APP23 mouse model of Alzheimer’s disease, brain regions known to be vulnerable to the disease showed early, region-specific disruption of diurnal transcription prior to substantial amyloid plaque deposition. That timing detail is the finding’s most important feature for drug discovery specifically: a signal detectable before visible pathology is exactly the kind of biomarker an early-intervention trial needs, since CNS trials frequently fail in part because treatment begins only after substantial, often irreversible neurodegeneration has already occurred.

A disease signature that shows up before the pathology does

A recent study used large-scale spatial transcriptomics to map 24-hour rhythmic gene transcription across cortical and subcortical regions of the mouse brain, finding marked regional differences in rhythmicity, including distinct oscillatory signatures across cortical areas and along the front-to-back axis of the brain.

In the APP23 mouse model of Alzheimer’s disease, brain regions known to be vulnerable to the disease showed early, region-specific disruption of diurnal transcription prior to substantial amyloid plaque deposition. That timing detail is the finding’s most important feature for drug discovery specifically: a signal detectable before visible pathology is exactly the kind of biomarker an early-intervention trial needs, since CNS trials frequently fail in part because treatment begins only after substantial, often irreversible neurodegeneration has already occurred.

That finding reframes what a spatial neurodegeneration biomarker program should actually look for. Rather than only mapping where existing pathology sits, the more valuable question is which region-specific molecular signals change first, before pathology becomes irreversible, since those early signals are what an early-intervention drug program actually needs to detect and act on.

Region-specific targets

Parkinson’s disease supplies a concrete, gene-level example of regional vulnerability, distinct from the circadian timing question above but addressing the same underlying principle: disease risk concentrates in specific brain regions, and understanding why requires knowing which genes are active in exactly those regions.

A Communications Biology study integrated Allen Human Brain Atlas gene expression data with post-mortem brain tissue characterized across Braak Lewy body stages, the standard pathological staging system for Parkinson’s disease progression. The analysis identified specific genes associated with regional vulnerability, including SCARB2, ELOVL7, SH3GL2, SNCA, BAP1, and ZNF184, several of which are already established Parkinson’s genetic risk factors, confirming the regional expression analysis was capturing biologically meaningful signal rather than incidental correlation.

Co-expression analysis across the vulnerable regions identified two distinct functional modules, both highly expressed specifically in regions involved in the disease’s preclinical stages: one enriched for genes related to dopamine synthesis and microglia, the other for genes related to the immune system, blood-oxygen transport, and endothelial cells. That second module is worth noting specifically, since it points toward vascular and immune contributions to regional vulnerability, a mechanism distinct from the dopaminergic neuron loss Parkinson’s disease is most commonly associated with.

Finding	Disease	Why it matters for drug discovery
Region-specific circadian disruption preceding plaque deposition	Alzheimer’s disease (mouse model)	Identifies a potential early biomarker detectable before irreversible pathology, relevant to early-intervention trial design
Named regional vulnerability genes and two functional modules	Parkinson’s disease	Points toward specific druggable pathways, including a vascular/immune module distinct from dopaminergic neuron loss, active in preclinical disease stages

Table 1. Two region-specific findings across two different neurodegenerative diseases, both pointing toward disease biology detectable before advanced, symptomatic pathology.

The neural microenvironment

Both findings above point toward the same broader principle: neurodegeneration is rarely a story about neurons alone. The Parkinson’s vulnerability modules specifically implicate microglia, immune signaling, and endothelial cells alongside dopaminergic neurons, and the Alzheimer’s circadian finding describes a brain-wide, multi-region phenomenon rather than a single cell type’s isolated dysfunction.

That pattern argues for treating the neural microenvironment, the specific mixture of neurons, glia, vasculature and immune cells present in a given brain region, as itself a unit of analysis, rather than studying neuronal dysfunction in isolation and treating supporting cell types as background. A drug candidate designed purely against a neuronal target may be addressing only one component of a regionally organized, multi-cell-type disease process that spatial data is what actually reveals as multi-component in the first place.

Model systems and translation

Even a well-characterized spatial signature discovered in a mouse model faces the same translation problem that has long plagued CNS drug discovery generally: rodent models of neurodegenerative disease are widely recognized as imperfect predictors of human disease biology and human drug response, a documented concern across Alzheimer’s, Parkinson’s, and other neurodegenerative disease research specifically.

Spatial biology does not solve that translation gap by itself, but it does supply a genuinely useful tool for narrowing it: a molecular signature identified spatially in a mouse model, such as the circadian disruption finding above, can be checked directly against human post-mortem tissue or brain atlas data, such as the Allen Human Brain Atlas resource the Parkinson’s study drew on, to test whether the same region-specific pattern actually holds in human tissue before a program invests further resources assuming it does.

The practical methodology for designing a brain spatial transcriptomics study, including platform selection, RNA integrity assessment, and postmortem tissue handling considerations specifically relevant to human brain material, is covered by our colleagues at Technology Networks in Spatial Transcriptomics in Neuroscience and Brain Atlases, a genuinely useful companion resource for the research-methodology side of a program this spoke addresses from the drug-discovery-strategy side.

For where these emerging therapeutic areas sit within spatial biology’s broader trajectory, see Where Spatial Biology is Headed in Drug Discovery, and for where this fits within the full spatial biology pipeline, Spatial Biology in Drug Discovery: From Target Discovery to Translational Medicine.

This article was produced under Drug Discovery News’s AI editorial policies.