Spatial multi-omics reveals region-specific molecular signatures in a 6-OHDA model of Parkinson’s disease Authors: Sun Young Lee, Hyun Kyong Shon, Amos Chungwon Lee, Young-Kyoung Ryu, Hye-Yeon Park, Kyoung-Shim Kim, Ga Seul Lee, Sunghoon Kwon, Chul Ho Lee, Jeong Hee Moon, Tae Geol Lee, Jin Gyeong Son Published: 2026 ABSTRACT AbstractParkinson’s disease (PD) is characterized by complex molecular and circuit-level alterations that extend beyond dopaminergic neurodegeneration, yet the spatial integration of metabolic and proteomic changes remains insufficiently explored. Here, we combined time-of-flight secondary ion mass spectrometry–based metabolite imaging with laser cell sorting proteomics to interrogate molecular alterations in the substantia nigra and striatum of the 6-hydroxydopamine toxin model of PD. Our analyses revealed distinct region-specific metabolic and proteomic signatures within primary lesion sites, and additionally uncovered unexpected and widespread off-target changes across neural circuits. These findings demonstrate that even in a toxin-induced model, PD pathology involves extensive reorganization of molecular networks and circuit-level processes, underscoring the complexity and diffuseness of disease mechanisms. By applying a spatially resolved, multilayered approach, this study expands the pathophysiological understanding of PD and provides a foundation for future investigations into the spatial and temporal dynamics of neurodegenerative disease progression.Subject terms: Biological techniques, Computational biology and bioinformatics, Neuroscience Introduction Parkinson’s disease (PD) is characterized by the progressive and selective degeneration of dopaminergic neurons in the substantia nigra pars compacta and the accumulation of Lewy bodies enriched in misfolded and aggregated α-synuclein1,2. While these canonical hallmarks define the disease, it is now widely accepted that PD represents a multifactorial disorder in which diverse pathogenic mechanisms, ranging from protein aggregation and impaired proteostasis to mitochondrial dysfunction, lysosomal impairment, and neuroinflammation, intersect within a vulnerable neuronal milieu3–5. Within this multifactorial landscape, metabolic dysregulation has emerged as a critical determinant of neuronal vulnerability, serving as a molecular bridge between organelle stress, disrupted signaling, and compromised membrane integrity6,7. From a metabolic perspective, accumulating evidence points to profound disturbances in lipid metabolism and energy homeostasis across both patient-derived samples and experimental models. Altered lipid composition has been linked to disrupted membrane architecture, abnormal vesicular trafficking, and compromised myelin integrity8–10. Lipidomics studies consistently report shifts in phosphatidylcholine (PC), phosphatidylethanolamine (PE), sphingolipids, and cholesterol derivatives11,12, molecules that directly influence membrane curvature, fluidity, and protein–lipid interactions. In parallel, proteomic investigations have revealed converging disruptions in mitochondrial respiratory chain complexes, autophagy–lysosomal proteins, and chaperone-mediated folding machinery13,14, which together highlight a systemic breakdown of cellular energy and proteostasis networks. Importantly, lipid remodeling does not represent a passive phenomenon but acts as an active modulator of α-synuclein aggregation, exerting its influence through preferential interactions with anionic membrane surfaces and by facilitating nucleation events that drive fibril assembly15. These findings indicate that disturbances in lipid homeostasis are tightly coupled to failures in protein quality control and collectively drive PD pathogenesis. However, because much of the current knowledge is still derived from bulk tissue analyzes, the spatial heterogeneity of metabolic and proteomic alterations across neuronal circuits remains insufficiently resolved. This fact underscores the need for approaches that preserve anatomical context. To address this gap, spatially resolved approaches offer the ability to situate molecular alterations within intact neuroanatomical circuits. One such approach is mass spectrometry imaging (MSI), which enables unbiased mapping of metabolites, lipids, and proteins directly in intact tissue sections. Early studies demonstrated the feasibility of extending MSI beyond small molecules to proteins, followed by the development of top-down proteomics and multimodal workflows such as matrix-assisted laser desorption/ionization mass spectrometry imaging (MALDI-MSI) combined with laser cell sorting16,17 and liquid chromatography–tandem mass spectrometry (LC-MS/MS)18–20. These advances provide a framework for constructing spatially informed molecular profiles of disease-relevant microenvironments and lay the foundation for integrative investigations into PD pathology. For example, MSI-based analyzes of PD brain tissue have visualized glutathione and related metabolic perturbations in degenerating nigral regions21 and more recently revealed abnormal processing of neuropeptides, such as substance P and dynorphins, within the basal ganglia22. These studies emphasize the utility of MSI in capturing spatially confined molecular alterations that are directly relevant to PD pathogenesis. In this study, we implemented a spatially guided multi-omics strategy that integrates lipid imaging via high-resolution time-of-flight secondary ion mass spectrometry (ToF-SIMS) with proteomics based on laser cell sorting. Using the unilateral 6-hydroxydopamine (6-OHDA) lesion model, which induces selective dopaminergic degeneration while preserving hemispheric symmetry, we were able to perform anatomically precise within-animal comparisons23,24. This integrative design is expected to reveal circuit-specific remodeling events that may otherwise be obscured in bulk analyzes, thereby providing new mechanistic insights into PD pathology. Our analysis first focused on the canonical nigrostriatal pathway (namely, that from the substantia nigra to the caudate–putamen, SN–CP), a motor circuit particularly vulnerable in early PD25,26. ToF-SIMS imaging revealed lipid remodeling that we could directly link to proteomic alterations through laser-cell-sorting-guided profiling. Beyond the SN–CP axis, we observed additional lipid shifts in adjacent white matter tracts, including fiber bundles connecting midbrain nuclei to higher circuits and major motor pathways descending from the cortex. These findings suggest that axonal and myelin-rich regions undergo parallel or secondary degeneration in response to nigral lesions. Regions with distinct lipid signatures—including both motor and non-motor domains—were subsequently isolated for LC-MS/MS–based proteomic analysis. Together, results demonstrate how the multimodal strategy delineates lipid–protein interplay within the canonical dopaminergic pathway while simultaneously uncovering underappreciated white matter vulnerabilities, helping to broaden our understanding of PD pathology and establish a framework for anatomically informed multi-omics investigations. Results Overview of the spatially resolved molecular findings Our integrated spatial analysis combining ToF-SIMS lipid imaging and LC-MS/MS proteomics revealed distinct molecular alterations across anatomically and functionally defined brain circuits in a 6-OHDA–induced PD mouse model (Fig. 1a). To establish this system, unilateral 6-OHDA lesions were generated to enable direct comparison of lesioned and intact hemispheres within the same animal. This framework provided a foundation for dissecting circuit-level biochemical vulnerability during dopaminergic neurodegeneration. Integrated spatial workflow for circuit-level molecular analysis in a 6-OHDA–induced Parkinson’s disease mouse model. a Unilateral 6-OHDA lesioning targeting the medial forebrain bundle (MFB) to establish a Parkinson’s disease model; the inset illustrates the mechanism of dopaminergic neuron apoptosis induced by reactive oxygen species (ROS) via the dopamine transporter (DAT), with molecular changes analyzed two weeks post-injection. b Coronal brain cryosectioning and sample preparation: glass slides for tyrosine hydroxylase (TH) immunohistochemical (IHC) staining to validate dopaminergic neuron loss in the substantia nigra (SN) and caudate–putamen (CP), stainless steel (SUS) slides for time-of-flight secondary ion mass spectrometry (ToF-SIMS) imaging, and indium tin oxide (ITO)-coated slides for laser cell sorting (LCS) and proteomic analysis. c ToF-SIMS imaging for untargeted spatial lipidomics, generating high-resolution ion distribution maps. d Region-of-interest (ROI) selection based on lipid signatures (indicated by red outlines), followed by precise dissection of the SN and CP subregions from adjacent ITO-mounted SN included brain tissue (SNiBT) and CP included brain tissue (CPiBT) sections using LCS. e Proteomic analysis of sorted ROIs with liquid chromatography–tandem mass spectrometry (LC-MS/MS) to directly link lipid remodeling to proteomic network alterations. To verify the extent of the lesions, brain tissue was cryosectioned and distributed onto three substrates (Fig. 1b). Glass slides were used for immunohistochemical (IHC) staining, stainless steel plates were used for ToF-SIMS imaging, and indium tin oxide (ITO)-coated slides were prepared for spatial protein analysis via the laser cell sorting technique16,17. This coordinated preparation ensured that histological validation, metabolomic imaging, and proteomics could all be pursued within the same experimental framework. Following preparation, ToF-SIMS lipid imaging provided high-resolution ion distribution maps that enabled untargeted profiling of regional biochemical changes (Fig. 1c). The lipid signatures guided the definition of regions of interest (ROIs) in the SN and CP, which were subsequently isolated by laser cell sorting from adjacent ITO-mounted sections (Fig. 1d). The anatomical accuracy of these ROIs was further validated through cross-referencing with Tyrosine Hydroxylase (TH) staining of adjacent sections, ensuring that the biochemically defined regions aligned with the extent of dopaminergic loss. The microdissected tissues were then analyzed by label-free spatial proteomics using nanoflow liquid chromatography–electrospray ionization–tandem mass spectrometry (nLC-ESI-MS/MS) (Fig. 1e). This workflow allowed for the region-matched integration of lipid remodeling and proteomic network alterations, even when feature-level correlations were divergent. Progressing from lesion validation to ROI-guided proteomics through spatial lipidomics, the developed integrative workflow offered a multilayered perspective on PD pathology by capturing both neurochemical and proteomic dimensions of dopaminergic circuit vulnerability. Spatial lipid remodeling revealed by MSI in response to 6-OHDA toxicity Our imaging approach was first directed at the SN–CP axis, the primary target of 6-OHDA toxicity and a canonical circuit central to motor control. Because dopaminergic neurons are highly enriched in membrane lipids, their selective loss is expected to trigger profound alterations in local lipid composition. Thus, ToF-SIMS imaging of the SN and CP was undertaken to capture these primary lipid perturbations in situ. The goal was to determine whether specific lipid classes exhibit region-specific spatial remodeling that reflects key pathological processes, such as oxidative stress, neuroinflammation, or membrane destabilization. Lipids were either identified from established ToF-SIMS lipidomic references or confirmed by nLC-ESI-MS/MS analysis (Supplementary Table S1 and Fig. S1). Negative-mode pooled LC-MS/MS analysis processed by MS-DIAL detected 188 aligned lipid-related features. Because many ToF-SIMS peaks could not be assigned confidently at the molecular-species level, the main manuscript focused on a subset of 13 lipid species that were supported by orthogonal LC-MS/MS analysis and could be interpreted with reasonable confidence. In the SN region of the lesioned hemisphere, ion images revealed clear elevations of fatty acid (FA) 16:1 and FA 18:1, while FA 16:0 appeared largely symmetric, FA 18:0 showed a modest reduction, and FA 20:4 exhibited a slightly decreased signal (Fig. 2a). Phospholipid imaging further showed localized increases in phosphatidylserine (PS) 36:1 and phosphatidylinositol (PI) species (36:3, 36:1), whereas other species such as PE 40:6 and PS 40:6 appeared more symmetric (Fig. 2c, e). Separately, triglyceride (TG) 52:2, a neutral lipid, also displayed localized elevation in the lesioned side. Beyond the targeted SN, we also analyzed adjacent regions within the same tissue (referred to here as SN included brain tissue, SNiBT) to assess whether 6-OHDA toxicity induces broader lipid remodeling. Effect size comparisons with the contralateral hemisphere revealed distinct and region-specific signatures. In the hippocampal region (SNH) of the lesioned hemisphere, most lipid signals were reduced, suggesting a generalized susceptibility (Fig. 2a). In contrast, the cortex (SNC) displayed selective lipid elevations, including significant increases in FA 16:1 and FA 16:0 as well as phospholipid remodeling marked by PS 36:1, TG 52:2, and PI species (36:3, 36:1) (Fig. 2c). These divergent responses indicate that 6-OHDA pathology extends beyond the primary dopaminergic axis, manifesting as diffuse signal loss in the hippocampus but focal lipid enrichment in the cortex, thus underscoring the region-specific nature of lipid remodeling in this model. These image-based observations were validated by effect size analysis across three biological replicates, summarized as a heatmap of Cohen’s d values (Fig. 2g). Because the number of confidently interpretable lipid species per ROI was limited after stringent assignment, lipid alterations were summarized primarily by effect size to emphasize reproducible directional differences between hemispheres rather than exhaustive ROI-level statistical inference. In the heatmap of the SNiBT section, the SN and SNC regions showed strong increases in FA 16:1 and FA 18:1 in the lesioned area as well as significant elevations in PS 36:1, TG 52:2, PI 36:3, and PI 36:1. Together, these results demonstrate that monounsaturated fatty acid (MUFA) enrichment (FA 16:1, 18:1) and remodeling of specific phospholipids are consistent features of the lesioned SN, while changes in saturated fatty acids (SFAs) and polyunsaturated fatty acids are comparatively subtle and heterogeneous. ToF-SIMS imaging and effect size analysis of lipid species in the SN and CP regions of mouse brain sections. Representative ion images are shown for selected lipid species, with m/z values and molecular annotations indicated above each column (see Table S1 for identification methods). For each region, the unlesioned (left) and 6-OHDA–lesioned (right) hemispheres are displayed side by side. a, c, e SN regions; b, d, f CP regions. g, h Heatmap visualizations of Cohen’s d effect sizes summarizing lipid distribution differences between lesioned and unlesioned hemispheres across three biological replicates (n = 3). Red indicates higher abundance in the lesioned hemisphere, and blue indicates lower abundance. In g, substantia nigra, SN; hippocampus, SNH; cortex, SNC. In h, caudate–putamen, CP; cortex, CPC; piriform, CPP. Next, considering its role in the canonical dopaminergic pathway, the CP was also included to assess whether lipid remodeling extended along this projection system. In the CP, lipid alterations were comparatively modest (Fig. 2b). Within the CP proper, effect size analysis across replicates revealed selective increases in PS 36:1 and PI 36:1, whereas most other species remained relatively unchanged (Fig. 2d). In the CP included brain tissue (CPiBT), the piriform cortex (CPP), however, displayed a broad reduction in lipid signals, with only FA 16:1, FA 16:0, and FA 18:1 showing modest increases (Fig. 2b). This widespread reduction in the CPP may provide a molecular correlate to the olfactory dysfunction recognized as an early non-motor symptom of PD27. By contrast, the cortical region of the CP (CPC) exhibited only minor elevations in FA 16:1, FA 16:0, and FA 18:1, changes that were not biologically substantial. The corresponding heatmap also showed the same results (Fig. 2h). Such findings suggest that while dopaminergic degeneration in the SN drives robust lipid remodeling locally and selectively extends to distal regions such as the SNH and CPP, the CPC remains relatively unaffected. Overall, these region-resolved lipidomic analyzes reveal that 6-OHDA–induced dopaminergic neurodegeneration is accompanied by a consistent enrichment of MUFAs (FA 16:1, FA 18:1) and selective remodeling of phospholipids within the SN, while more modest and heterogeneous changes occur in anatomically connected or distal regions. The hippocampus exhibited a generalized reduction in lipid signals, the cortex showed focal elevations in fatty acids and phospholipids, and the piriform cortex demonstrated widespread lipid depletion consistent with early olfactory dysfunction. By contrast, the CPC cortex remained largely unaffected. This pattern underscores that lipid remodeling in PD models is both region-specific and pathway-dependent, reflecting a combination of local oxidative stress, membrane perturbation, and circuit-level vulnerability28. The region-dependent lipid alterations suggest possible circuit-level effects of dopaminergic degeneration, which were further investigated by proteomic profiling. Qualitative assessment of spatial proteomic analysis To establish a foundation for downstream functional interpretation, we assessed the region-specific proteome of brain tissues from a 6-OHDA mouse model. Region-specific proteome analysis was performed after isolating ROIs from mouse brain tissues using laser cell sorting technology. Based on the ToF-SIMS imaging results, the ROIs were selected according to differences in lipid expression between lesioned and unlesioned regions. As shown in Fig. 3a, the SN, SNC, and SNH were selected as ROIs in the SNiBT section, while as shown in Fig. 3b, the CP, CPC, and CPP regions were selected in the CPiBT section. Unlesioned and lesioned hemispheres are labeled with L and R, respectively, in the images. Proteomic profiling of SNiBT and CPiBT sections from the 6-OHDA mouse model sorted by laser cell sorting. Representative images of a SNiBT (substantia nigra, SN; hippocampus, SNH; cortex, SNC) and b CPiBT (caudate–putamen, CP; cortex, CPC; piriform, CPP) regions defined based on lipidomic ion maps. Red outlines indicate regions of interest (ROIs) dissected for proteomic analysis. L and R indicate the unlesioned and lesioned hemispheres, respectively, separated by a dashed line. Proteomic results for the c SNiBT and d CPiBT sections. Bar graphs represent the number of identified protein groups per sample (left y-axis), and line graphs indicate the identified peptide counts (right y-axis). Proteomic analysis of tissue samples from six mice demonstrated consistent and reproducible protein expression patterns across all ROIs. In Fig. 3c, d, the bar graphs indicate the number of identified protein groups, and the line graphs (black circles) denote the peptide counts. In the SNiBT section (Fig. 3c), approximately 2000–3000 protein groups and 10,000–20,000 peptides were identified per ROI. In the CPiBT section (Fig. 3d), approximately 2000–4000 protein groups and 10,000–30,000 peptides were identified per ROI. Although detection rates varied across tissue regions, the numbers of protein groups and peptides detected within each group were consistent across biological replicates and regions, suggesting high technical reproducibility and sufficient input for subsequent analysis. These data confirm the robustness of our spatial proteomics workflow and validate the quality of tissue acquisition and processing across brain regions. Region-specific proteomic alterations Brain tissue samples of the SNiBT section, including the SN, SNC, and SNH, were subjected to proteomic profiling. Volcano plots revealed region-specific proteins based on criteria of a 1.5-fold change and p-value threshold of 0.05. In the SN, 45 downregulated and 6 upregulated proteins were identified (Fig. 4a), while in the SNC, 18 proteins were downregulated and 8 proteins were upregulated (Fig. 4b). By contrast, the SNH exhibited 19 downregulated and 222 upregulated proteins (Fig. 4c). GO term enrichment analysis was then performed to identify the major biological processes associated with the DEPs, with interpretation restricted to the most biologically coherent terms, particularly in regions with relatively small DEP sets. In the SN, downregulated proteins were associated with mitochondrial ATP synthesis, oxidative phosphorylation, the electron transport chain, and cellular respiration (Fig. 4d), whereas upregulated proteins were enriched in amino acid catabolic processes such as tyrosine, phenylalanine, lysine, and aromatic amino acid metabolism (Fig. 4e). In the SNC, downregulated proteins were enriched in oxidative phosphorylation, ATP metabolism, and reactive oxygen species regulation (Fig. 4f). In the SNH, downregulated proteins were linked to synaptic vesicle transport, regulation of vesicle fusion, and neurotransmitter release (Fig. 4g), while upregulated proteins were enriched in cellular protein localization, protein complex organization, and membrane-associated transport (Fig. 4h). Across all three regions, a common pattern was the enrichment of downregulated proteins in oxidative phosphorylation–related pathways, indicating a shared vulnerability of energy metabolism. However, the SNH displayed a distinct trend characterized by large-scale protein upregulation, particularly in processes related to protein localization and intracellular transport. This highlights region-specific molecular signatures within the SN. Proteomic analysis between lesioned and unlesioned areas in the SNiBT section of 6-OHDA mouse models. Volcano plots of proteins obtained from the a SN, b SNC, and c SNH regions. Red dotted lines indicate p = 0.05 and ±1.5-fold change thresholds. Proteins highlighted in blue are downregulated, while those in pink are upregulated. Biological process (BP) analysis of downregulated (d, f, g) and upregulated (e, h) proteins in the SN, SNC, and SNH regions. Significantly enriched BPs (p < 0.05) are presented. The bubble size represents the number of genes, and the color indicates the −log₁₀(FDR). Data were analyzed using the ShinyGO bioinformatics tool (version 0.82). Figure 5 details the proteomic alterations in the CP, CPC, and CPP regions in the CPiBT section, with the volcano plots demonstrating region-specific differences in protein abundance (Fig. 5a–c). In the CP, 7 proteins were downregulated, and 4 were upregulated, and in the CPC, 11 proteins were downregulated, and 7 were upregulated. In the CPP, 7 proteins were downregulated, and 134 were upregulated. Enrichment analyzes further characterized these changes. In regions with smaller DEP pools, these GO categories should be interpreted cautiously, as they may reflect broader adaptive or secondary responses rather than primary PD-driving mechanisms. Proteomic analysis between lesioned and unlesioned areas in the CPiBT section of 6-OHDA mouse models. Volcano plots of proteins obtained from the a CP, b CPC, and c CPP regions. Red dotted lines indicate p = 0.05 and ±1.5-fold change thresholds. Proteins highlighted in blue are downregulated, and those in pink are upregulated. Biological process (BP) analysis of upregulated (d, f, h) and downregulated (e, g) proteins in the CP, CPC, and CPP regions. Significantly enriched BPs (p < 0.05) are presented. Bubble size represents the number of genes, and the color indicates the −log₁₀(FDR). Data were analyzed using the ShinyGO bioinformatics tool (version 0.82). We then likewise investigated the biological processes associated with the downregulated proteins in the CPiBT section. In the CPC region, downregulated proteins were involved in nuclear protein localization, calcium ion membrane transport, and organelle organization (Fig. 5e), whereas proteins downregulated in the CPP region were linked to skeletal muscle contraction, glucosamine metabolism, and dendritic transport (Fig. 5g). Overall, proteins downregulated in the CPiBT section were commonly associated with essential cellular functions, with significantly reduced expression levels observed in the damaged brain regions. Region-specific differences revealed that proteins downregulated in the CPC region were primarily involved in cellular positioning processes, which could lead to dysfunction or long-term cellular damage. In contrast, proteins downregulated in the CPP region were predominantly associated with metabolic functions. Disruption of these metabolic pathways can impair cellular energy supply, potentially leading to long-term consequences on brain cell physiology29. In terms of upregulation, proteins upregulated in the CP region were involved in cardiac conduction, cell adhesion, and systemic process regulation (Fig. 5d). In the CPC region, upregulated proteins were associated with sepiapterin metabolism, glial differentiation, and central nervous system development (Fig. 5f), while those upregulated in the CPP region were implicated in long-term synaptic plasticity, neurotransmitter receptor signaling, and nervous system development (Fig. 5h). Upregulated proteins in each region were concentrated in pathways related to the cell cycle and division (CPC), neuroprotection and survival (CPP), and energy metabolism and transport (CP), reflecting region-specific responses that align with cellular repair mechanisms. Discussion Our central question was whether 6-OHDA–induced neurodegeneration leads to localized versus systemic molecular responses in the nigrostriatal system. By combining spatial lipidomics with region-specific proteomics, we were able to dissect whether lipid remodeling and proteomic alterations occur in parallel or diverge across anatomically distinct regions. We first examined the SN, the primary lesion site in PD models, before extending our analysis to downstream projection areas, including the CP. To examine the relationship between lipid MSI and proteomics, we performed matched-sample Spearman correlation analysis (n = 3). Direct feature-level correlations were modest (Supplementary Fig. S2), reflecting the distinct molecular dynamics of each omics layer. Therefore, we interpreted the cross-omics relationship primarily through functional convergence rather than strict molecule-to-molecule correspondence. In this integrated framework (Fig. 6), lipid MSI served as a spatial map to identify lesion-associated asymmetry, while proteomics of the corresponding regions characterized the mechanistic molecular programs associated with those lesion-defined areas. Integrated spatial multi-omics workflow and functional convergence in the 6-OHDA mouse model. a Lipid MSI reveals brain lipid asymmetry, characterized by altered fatty acid (FA) distribution and local membrane-state shifts, signaling region-selective tissue remodeling. b MSI-guided proteomics of SNiBT and CPiBT sections. Representative up- and down-regulated proteins, including APOE, ATP8A1, PLD3, FASN, FABP5/7, NPC2, and GPX4, correlate with spatial lipid remodeling. c Integration of findings reveals a circuit-wide continuum of remodeling. The neurodegenerative process involves mitochondrial failure and lipid dysregulation in the SN (local stress), organelle and transport remodeling in the SNH (structural reorganization), and plasticity and signaling compensation in the CPP (synaptic adaptation). In the SN, we observed clear MUFA enrichment (FA 16:1, FA 18:1) and selective PS/PI remodeling, together with proteomic downregulation of mitochondrial oxidative phosphorylation proteins (Figs. 2 and 3). This convergence reinforces established models that link mitochondrial dysfunction, oxidative stress, and membrane destabilization in dopaminergic neurons30–32. While the localized nature of these changes in the SN might raise concerns regarding potential injection-related artifacts, several lines of evidence support 6-OHDA-induced neurotoxicity as the primary driver. The 6-OHDA model is a well-established surrogate for selective dopaminergic loss, with pathological outcomes extending far beyond procedural trauma23. Critically, the distinct bilateral asymmetry observed in our analysis—where MUFA enrichment at the lesion site was accompanied by selective depletion of FA 18:0 and FA 20:4 relative to the contralateral SN—indicates specific metabolic reprogramming rather than non-specific mechanical trauma. This interpretation is further supported by in vitro literature demonstrating that 6-OHDA directly triggers lipidomic remodeling independent of physical stimuli33 and aligns with clinical evidence showing that lipid-rich membrane fragments are major constituents of Lewy pathology8,11. Having established the pathological specificity of these localized changes, we further found that molecular signatures varied significantly across interconnected brain areas, revealing a spatial gradient of proteomic adaptation. In both the SN and SNC, downregulation predominated, largely involving mitochondrial oxidative phosphorylation, ATP metabolism, and oxidative stress–related pathways34–36. These alterations are consistent with progressive energetic failure underlying dopaminergic vulnerability. In contrast, the hippocampal projection territory (SNH) exhibited the most extensive proteomic alterations, with 222 upregulated and 19 downregulated proteins (Fig. 4). To capture these biologically meaningful but moderate proteomic alterations, we utilized a 1.5-fold change threshold combined with a p-value < 0.05 and stringent peptide-level filters. While a more conservative twofold cutoff is sometimes used, our approach was optimized to preserve the sensitivity needed to detect coordinated pathway-level remodeling, such as the structural adaptations observed in the SNH, while maintaining high confidence through multi-parameter filtering. Pathways related to the synaptic vesicle cycle and vesicle fusion were markedly downregulated, indicating impaired synaptic transmission capacity37,38, while proteins involved in localization, organelle organization, and intracellular transport were strongly upregulated, suggesting a compensatory remodeling response39. Taken together, these data delineate a hierarchical continuum of molecular changes along the nigrostriatal axis: mitochondrial dysfunction originates locally in the SN, whereas broader structural remodeling programs are activated in secondary projection areas such as the SNH to preserve network connectivity and synaptic integrity under chronic stress40. Unlike in the SN, lipid alterations in the CP core were relatively limited, in line with modest direct pathology at this projection site41. By contrast, the CPP exhibited a striking upregulation of proteins associated with synaptic plasticity and neurotransmitter signaling pathways, despite minimal lipid remodeling42. These molecular signatures extend previous observations that partial nigral degeneration is sufficient to perturb olfactory-associated striatal circuits43. Indeed, mouse studies have shown that even mild dopaminergic neuron loss in the SN leads to measurable olfactory discrimination and memory impairments, along with altered odor-evoked activity in the olfactory bulb44. Concordantly, α-synuclein transgenic models have revealed inclusions spanning both SN and olfactory bulb, accompanied by reduced striatal dopamine and defective olfactory behaviors45. Collectively, our CPP findings indicate that nigral pathology not only produces local mitochondrial–lipid stress but also propagates into olfactory-linked striatal territories, thereby providing a mechanistic framework for the prodromal hyposmia frequently observed in PD45. Results of the parallel analyzes of SN–SNH and CP–CPP circuits highlight that our initial focus on nigrostriatal projection pathology only captures part of the picture. Beyond the expected convergence of mitochondrial dysfunction and lipid remodeling within the SN core, we observed that proteomic remodeling was often more pronounced in trajectories outside the primary target regions. In both the hippocampal and olfactory-linked striatal territories, large-scale alterations in synaptic and organelle-related proteins emerged in the relative absence of extensive lipid changes. This dissociation indicates that projection sites commonly regarded as secondary can undergo early proteomic change, suggesting their value as sensitive readouts of disease-associated stress. This spatial divergence further explains the modest feature-level correlation observed across the entire dataset, as the biological response of each omics layer varies depending on the anatomical context. By integrating lipidomic and proteomic profiling across interconnected regions, our approach was able to delineate canonical nigral alterations as well as uncover unexpected remodeling in off-target projection territories. These findings support the utility of spatially resolved, multimodal profiling as a strategy to detect prodromal circuit vulnerability and to broaden the mechanistic understanding of PD beyond the classical nigrostriatal axis. From an integrative perspective, these datasets reveal a circuit-wide continuum of molecular remodeling across the nigrostriatal pathway (Fig. 6). Specifically, our findings suggest a progressive pathological transition: starting from local oxidative stress and lipid dysregulation in the SN, moving toward structural reorganization in the SNH, and ultimately leading to synaptic adaptation and signaling compensation in the CPP. This spatial-mechanistic continuity demonstrates that while feature-level concordance may be divergent, the two omics layers synergistically converge on unified pathological programs. Accordingly, the major value of this integration lies in capturing these complementary and systemic aspects of lesion biology, providing a more holistic understanding of the 6-OHDA model beyond individual molecular changes. The present study should also be interpreted in light of certain methodological considerations. The analysis was structured around lesion-centered molecular changes in the substantia nigra and their extension to anatomically connected regions. In this context, the observed SN-associated remodeling may also have implications for broader motor circuitry, although neighboring midbrain structures were not separately quantified in the current study. A limitation of the present imaging-based lipid analysis is that, although it provides strong spatial information, the number of confidently assigned lipid species per ROI remained limited for extensive statistical inference. We therefore used effect sizes to summarize robust directional changes, while recognizing that deeper orthogonal lipidomics would further strengthen quantitative validation. Furthermore, GO enrichment results from regions with smaller DEP sets should be interpreted with caution. In these cases, enriched categories may reflect secondary adaptive or compensatory responses rather than primary PD-driving mechanisms. We therefore focused our interpretation on the most biologically coherent themes that were consistent with the lesion-centered framework of the study, including mitochondrial dysfunction in the SN-related regions and adaptive intracellular transport or synaptic remodeling in connected regions. Additionally, although the contralateral hemisphere served as a useful intra-animal reference to reduce inter-animal variability—a factor of particular importance in high-sensitivity spatial multi-omics—it may not represent a completely unaffected baseline in the unilateral 6-OHDA model. Potential bilateral or systemic effects, including compensatory responses, may influence the contralateral side and thereby cause the observed molecular differences to reflect relative rather than absolute pathological changes. Future studies incorporating sham-operated (e.g., saline-injected) controls will be important for defining disease-associated alterations against a more strictly physiological baseline and to further separate toxin-driven changes from procedure-related local effects. Ultimately, the major value of this integration lies in capturing these complementary and systemic aspects of lesion biology. By bridging localized nigral injury to system-wide circuit remodeling, this approach provides an understanding of the 6-OHDA model beyond individual molecular changes, establishing a framework for future spatially-resolved investigations into neurodegenerative progression. In summary, our integrated ToF-SIMS–based lipid imaging and LC-MS/MS–based proteomic profiling have characterized circuit-specific alterations that transcend focal dopaminergic neuronal loss. The observation of unexpected and widespread off-target changes—even in a toxin-induced model—highlights that Parkinsonian pathology involves a profound reconfiguration of neural circuits and molecular networks. By revealing circuit–molecule relationships that may be overlooked by conventional bulk analyzes, this spatially resolved strategy expands the fundamental understanding of PD pathophysiology. These findings establish a foundation for future longitudinal studies incorporating human subjects to further delineate the temporal and spatial dynamics of disease progression. Methods Materials Bicinchoninic acid (BCA) assay reagent, n-dodecyl beta-D-maltoside (DDM), acetonitrile (ACN), and water containing LC-MS grade 0.1% formic acid, trypsin, and SYBR gold nucleic acid stain dye were purchased from Thermo Fisher Scientific (Rockford, IL, USA). Gold chloride hydrate, sodium citrate tribasic dehydrate, tris(2-carboxyethyl) phosphine hydrochloride (TCEP), magnesium chloride, potassium cyanide, thiazolyl blue tetrazolium bromide, hydrochloric acid, propidium iodide, ammonium bicarbonate (AmBic), dithiothreitol (DTT), iodoacetamide (IAA), L-cysteine, sodium chloride (NaCl), formic acid, tris-HCl buffer (pH 8.0), and phosphate buffered saline (PBS) were purchased from Sigma-Aldrich (St. Louis, MO, USA). Protease inhibitor cocktail was obtained from Roche Diagnostic GmbH (Mannheim, Germany). To purify protein digests, HLB cartridges were purchased from Waters (MA, USA), and HPLC-grade water and ACN were purchased from Honeywell (MI, USA). Acetic acid and dimethyl sulfoxide were purchased from Samchun (Rep. of Korea). 6-hydroxydopamine (6-OHDA) and desipramine hydrochloride (25 mg/kg, intraperitoneally [i.p.]) were purchased from Sigma-Aldrich (St. Louis, MO, USA). 6-OHDA was dissolved in 0.2% ascorbic acid in saline, and desipramine was dissolved in saline. 6-OHDA lesion induction C57BL/6J mice were purchased from Jackson Laboratory (Bar Harbor, ME, USA) and maintained at the Korea Research Institute of Bioscience and Biotechnology (KIRBB, Daejeon, Korea). The mice were housed in polycarbonate plastic cages in a temperature (22 ± 2 °C) and humidity (50–60%) controlled environment and specific-pathogen-free conditions under a 12 h light/dark cycle (lights on at 7 am). The mice had free access to food and water. Mice were handled in accordance with the Guidelines of Animal Care of the Korea Research Institute of Bioscience and Biotechnology (KRIBB-AEC-20144). The procedure for nigral 6-OHDA lesion-induction has been previously described46,47. The mice were pretreated with desipramine hydrochloride 30 min prior to surgery to prevent noradrenergic damage, and then were anesthetized with a combination of ketamine hydrochloride and xylazine hydrochloride. Each mouse received a unilateral injection of 15 μg (5 μg/μl) 6-OHDA into the left substantia nigra pars compacta using the following coordinates from the Paxinos and Franklin mouse brain atlas (2007): anteroposterior (AP), −3.0 mm; mediolateral (ML), −1.3 mm; and dorsoventral (DV), −4.7 mm. Following surgery, mice were kept on a warming plate until full recovery from anesthesia and were then returned to their home cages. To prevent dehydration, glucose in saline (10 ml/kg, s.c.) was administered. Tissue preparation and cryosectioning Two weeks after 6-OHDA lesioning, mice were sacrificed and their brains carefully removed. The extracted brains were immediately immersed in 2-methylbutane for 30 min and stored at −80 °C to minimize ice crystal formation. For sectioning, brains were equilibrated in a cryostat chamber (Leica CM3050S) for 30 min, embedded in OCT, and cut into consecutive 12 μm coronal sections. Sections were mounted on stainless steel plates for ToF-SIMS, ITO-coated slides for LCS, or standard glass slides for IHC. All samples were freeze-dried under vacuum for 3 h before analysis. Immunohistochemistry Frozen brain sections (12 µm) mounted on glass slides were stored at −80 °C until use. For IHC staining, slides were treated with 30% hydrogen peroxide to block endogenous peroxidase activity. After washing with PBS, sections were incubated with 5% horse serum for 1 h at room temperature to block nonspecific binding, followed by incubation with the primary antibody against tyrosine hydroxylase (1:1000 in 5% horse serum, TH, Pel-freez, Rogers, AR, USA, P40101-0) overnight at 4 °C. The following day, sections were washed with PBS and incubated with a biotinylated anti-rabbit IgG secondary antibody (1:200 in 5% horse serum, Vector Laboratories, Inc., Burlingame, CA, USA) for 1 h at room temperature. For visualization, the avidin–biotin peroxidase complex (ABC kit, Vector Laboratories, Inc.) method was applied, and 3,3′-diaminobenzidine (DAB, Sigma-Aldrich) was used as the chromogen. After color development, slides were dehydrated through graded alcohols and mounted with balsam (Sigma-Aldrich). Images of the striatum and substantia nigra pars compacta were captured at 40× magnification (detailed in Supplementary Fig. S3). ToF-SIMS measurement The ToF-SIMS analysis was performed using a ToF-SIMS V (IONTOF GmbH) equipment. Mass spectra and images were obtained by using an argon cluster ion beam (Ar3000+, 20 keV) with a 0.15 pA beam current. Delayed extraction modes were used in all analyzes to improve mass resolution. Data were acquired in positive and negative modes from the same area within static limits using a primary ion dose density of approximately 2.5 × 1011 ions/cm2 for each polarity. Irganox 1010 was used for external mass calibration, where Irganox 1010 dissolved in acetone was dropped near the sample and dried. The mouse brain tissue sample was measured over a wide area using a stage scan, including the Irganox 1010 spot. To obtain images of the whole mouse brain tissue, a patch size of 30 μm and a stage raster size of 10,200 × 6210 μm2 were used. For external mass calibration, using the ROI (region of interest) function, mass spectra from the Irganox 1010 region were obtained. Quantification of ROI areas To ensure the reproducibility of the microdissection process, the actual area of each ROI was quantified using pixel-based image analysis. For representative samples (n = 3), the total pixel counts of the designated ROIs were converted into absolute areas (mm2) based on the image resolution. This validation confirmed that the sampling volumes remained consistent across both hemispheres and biological replicates, minimizing potential quantitative bias in the subsequent proteomic analysis (detailed in Supplementary Table S2 and Fig. S4). Lipid extraction Lipids were extracted from LCS-dissected brain tissues using a scaled-down MTBE-based protocol adapted from a previous report48. The extraction was initiated with 200 µl of MTBE to accommodate the limited tissue amounts. After phase separation, the organic layer was collected, dried under vacuum, and stored at −80 °C until analysis. LC-MS/MS analysis for lipidomic analysis LC-MS/MS was performed on an LTQ-Orbitrap Elite (Thermo Fisher Scientific) equipped with a nanoAcquity system (Waters). Lipids were separated on an Atlantis dC18 column (3 µm, 300 µm × 100 mm P/N186003505; Waters) at 50 °C at a flow rate of 4 µl/min. The mobile phases were 60:40 acetonitrile/water (buffer A) and 90:10 isopropyl alcohol/acetonitrile (buffer B); both A and B contained 10 mM ammonium formate and 0.1% formic acid. Dried lipid extracts were resuspended with 30 µl of buffer A, and 5 µl of the solution was injected. The separation was conducted under the following step gradient: from 32% B to 97% B in 21 min. Full scans were set as follows: spray voltage 1900 V, resolution 60,000 at m/z 400, scan range m/z 100–1000, positive mode. The MS/MS scans were set as follows: number of dependent scans, 5; threshold 10,000, isolation width 1, normalized collision energy 30, and dynamic exclusion (exclude after one time, exclusion duration 30 s, mass tolerance ±15 ppm). Raw data files were processed using MS-DIAL software (v4.9)49. The following parameters were used: mass accuracy MS1 0.01 Da and MS2 0.05 Da, minimum peak height 1000 amplitude, minimum peak width 5. Laser cell sorting from tissues and protein extraction Region-specific tissue samples were obtained using a pulsed-laser-based cell sorting system and subjected to protein extraction and digestion, as previously reported16,17. ROIs corresponding to predefined regions were isolated using a near-infrared pulsed laser cell sorting system. The identification and selection of these ROIs were primarily guided by ToF-SIMS lipid MSI maps, which provided high-resolution biochemical markers of lesion-associated remodeling. This molecularly-driven approach allowed for the isolation of altered tissue regions in their native state, preserving the maximum proteomic depth that could otherwise be compromised by the fixation and washing steps required for conventional immunohistochemical staining. To ensure anatomical accuracy and validate the 6-OHDA-induced lesion, adjacent tissue sections were subjected to Tyrosine Hydroxylase (TH) staining as a spatial reference. Individual cell clusters or single-cell regions were ejected into low-bind microcentrifuge tubes preloaded with 5 µL of LC-MS–grade lysis buffer. To minimize contamination, the collection device and tubes were prerinsed with LC-MS–grade water and 70% ethanol, and the laser path was cleaned between each collection. Protein extraction and digestion Collected tissues were lysed in 0.2% DDM, 150 mM NaCl, and 50 mM Tris–HCl (pH 7.5), supplemented with protease inhibitors. Samples were heated at 95 °C for 10 min with shaking and centrifuged at 20,000 × g for 30 min at 4 °C. Supernatants were collected, and the protein concentrations were determined by the BCA assay. For digestion, 20 µg protein was reduced with 10 mM DTT in 50 mM AmBic at 47 °C for 1 h, followed by alkylation with 20 mM iodoacetamide at room temperature for 30 min in the dark. Proteins were digested with sequencing-grade trypsin at a 20:1 (protein:enzyme, w/w) ratio for 18 h at 37 °C. Peptides were purified using HLB cartridges, dried under vacuum, and stored at −20 °C until LC-MS/MS analysis. nLC-ESI-MS/MS for proteomic analysis The dried samples were reconstituted in an aqueous solution containing 2% ACN and 0.1% FA (LC-MS grade). A 250 ng aliquot of the resulting peptide mixture was injected into a NanoElute LC system connected to a hybrid trapped ion mobility spectrometry–quadrupole time-of-flight mass spectrometer (timsTOF Pro, Bruker Daltonics, Bremen, Germany), which was equipped with a modified nano-electrospray ion source (CaptiveSpray, Bruker Daltonics). Peptides were separated on a handmade column (75 μm inner diameter, 250 mm length) packed with C18 resin (1.9 μm, 120 Å, Dr. Maisch, Germany) at 50 °C with a constant flow rate of 400 nl/min. Elution was performed using a binary gradient of mobile phases A (0.1% FA in H2O) and B (0.1% FA in ACN) as follows: 2% to 17% B for 45.0 min, 17% to 25% B for 22.5 min, 25% to 37% B for 7.5 min, 37% to 80% B for 5.0 min, and held at 80% B for 10 min to rinse the column. The timsTOF Pro mass spectrometer was operated in parallel accumulation–serial fragmentation mode using Bruker Compass HyStar 6.3.1.8. The MS and MS/MS settings were as follows: mass range m/z 100 to 1700, 1/K0 start at 0.6 Vs/cm² and end at 1.6 Vs/cm², capillary voltage of 1500 V, dry gas flow rate of 3 l/min, and dry temperature of 180 °C. For parallel accumulation serial fragmentation mode, 10 MS/MS scans were performed (total cycle time: 1.16 s), with a charge range of 0 to 5, active exclusion for 0.4 min, scheduling target intensity of 20,000, and an intensity threshold of 2500, depending on the precursor mass and charge. Data analysis The obtained raw data were analyzed using PEAKS Studio 10.5 (Bioinformatics Solutions, Waterloo, Canada) for protein identification and label-free quantification (LFQ) of cell-derived proteins. Searches were performed against the Swiss-Prot databases of Mus musculus (mouse, UP00000589) from UniProt (www.uniprot.org/) with a false discovery rate (FDR) of 0.01. The parameters for protein identification were as follows: (a) trypsin as the specific enzyme with up to two missed cleavages allowed; (b) fixed modification: carbamidomethylation of cysteine; variable modifications: oxidation of methionine, allowing for three variable post-translational modifications per peptide; (c) precursor mass error tolerance of 20.0 ppm; (d) fragment mass error tolerance of 0.05 Da. After protein identification, LFQ analysis was performed based on the PEAKS dataset. LFQ was conducted using the analysis of variance method, with protein identification controlled at <1% FDR at the peptide-spectrum match level in PEAKS Studio 10.5. To ensure the robust identification of differentially expressed proteins (DEPs) while balancing statistical sensitivity, we employed a multi-parameter filtering strategy rather than a separate Benjamini–Hochberg p-value adjustment. Specifically, proteins were defined as DEPs if they met the following criteria: a PEAKS significance score ≥13 (equivalent to p < 0.05), a fold-change threshold of ≥ 1.5 or ≤ 0.67, and a requirement of at least two unique peptides for quantification. Total ion current was employed for data normalization. Regarding the handling of missing values, we adopted a strategy to maximize the discovery of region-specific alterations while maintaining data integrity. In each experimental group (n = 6), only proteins detected in at least two out of six biological replicates (≥ 33.3%) were included in the downstream analysis. This detection threshold was selected to preserve potentially critical but low-abundance biological signals that might be significantly depleted in the 6-OHDA-lesioned regions. No global imputation was performed to avoid the introduction of artificial variance; instead, strict statistical filters and quality controls were applied to ensure that the reported DEPs represented robust biological shifts rather than stochastic noise. Gene Ontology (GO) term enrichment analysis was subsequently performed using the ShinyGO program (version 0.82) after exporting the data to Microsoft Excel. A comprehensive list of abbreviations and symbols used throughout this study is provided in Supplementary Table S3. Supplementary information Supplementary information Acknowledgements This work was supported by the Development of Measurement Standards and Technology for Biomaterials and Medical Convergence program funded by the Korea Research Institute of Standards and Science (No. KRISS-GP2025-0007-02), the National Research Council of Science & Technology (NST) (No. GTL24022-000), and the Nano & Material Technology Development Program (RS-2024-00452934) through the National Research Foundation of Korea (NRF) funded by the Ministry of Science and ICT, Republic of Korea. This research was also supported by a grant of the Korea-US Collaborative Research Fund (KUCRF), funded by the Ministry of Science and ICT and Ministry of Health & Welfare, Republic of Korea (RS-2024-00468338 to S.K., RS-2023-00266110 to A.C.L, NRF-2020R1A3B3079653 to S.K., RS-2024-00468338 to S.K. and A.C.L., and RS-2024-00454407 to S.L); and by the Industrial Strategic Technology Development Program (20024391, RS-2024-00451981 to A.C.L., RS-2024-00508416 to S.L) funded by the Ministry of Trade, Industry and Energy (MOTIE, Korea). Author contributions S.Y. Lee and H.K. Son contributed equally to this work. S.Y. Lee, H.K. Son, T.G. Lee, and J.G. Son designed the research. S.Y. Lee, H.K. Son, A.C. Lee, Y.K. Ryu, H.-Y. Park, K.-S. Kim, G.S. Lee, S. Kwon, C.H. Lee, and J.H. Moon performed the experiments. S.Y. Lee and H.K. Son wrote the original draft and all the authors discussed and commented on it. T.G. Lee and J.G. Son supervised and administered the project. Data availability The raw imaging MS datasets are available in the Korea Bio-Data Station (K-BDS) under accession number KAP242352. The proteomic data for this study are available in the PRIDE database under accession number PXD071834. The LC-MS/MS data used for lipid identification are available from the corresponding author upon reasonable request. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. These authors contributed equally: Sun Young Lee, Hyun Kyong Shon. Contributor Information Tae Geol Lee, Email: tglee@krissre.kr. Jin Gyeong Son, Email: yeskyoung@kriss.re.kr. Supplementary information The online version contains supplementary material available at https://doi.org/10.1038/s41531-026-01433-5. References Associated Data Supplementary Materials Supplementary information Data Availability Statement The raw imaging MS datasets are available in the Korea Bio-Data Station (K-BDS) under accession number KAP242352. The proteomic data for this study are available in the PRIDE database under accession number PXD071834. The LC-MS/MS data used for lipid identification are available from the corresponding author upon reasonable request.