NanoString GeoMx DSP Spatial Proteomics Service
CD Genomics provides an end-to-end NanoString GeoMx DSP Spatial Proteomics Service for ROI-based, high-plex protein profiling directly from FFPE or fresh-frozen tissue sections. The GeoMx Digital Spatial Profiler (DSP) uses UV-cleavable, oligo-barcoded antibodies to quantify protein expression in user-selected regions of interest (ROIs) — tumor nests, immune infiltrates, stromal compartments, or tertiary lymphoid structures — rather than averaging signal across the entire tissue section. Barcode counts are read out by NGS, decoupling detection from fluorescence channel limitations and enabling simultaneous measurement of protein targets at scale. The platform is especially valued for FFPE cohort studies where archived specimens, tissue heterogeneity, and spatial biomarker validation are central concerns.
- ROI-driven spatial proteomics: choose which tissue regions to profile based on histology or IF morphology
- High-plex protein detection with barcoded antibodies read by NGS — no spectral overlap constraints
- Core strength in FFPE tissue; also compatible with fresh-frozen sections
- Modular protein panels: immune cell typing, immune activation, IO drug targets, tumor signaling, and custom targets
Technology Overview: How GeoMx DSP Spatial Proteomics Works
The GeoMx DSP Platform
GeoMx DSP (originally developed by NanoString, now under Bruker Spatial Biology) is a spatial multi-omics platform that combines fluorescence-guided ROI selection with oligo-barcoded detection reagents read by next-generation sequencing. Unlike imaging-based spatial proteomics platforms that detect proteins via repeated cycles of fluorophore-labeled antibodies, DSP uses a single incubation with a cocktail of DNA-barcoded antibodies, then releases the barcodes from selected regions using focused UV light. The collected barcodes are quantified by NGS, where each barcode count corresponds to the abundance of a specific protein target.
This NGS readout is the defining technical advantage of DSP proteomics: protein detection is not limited by the number of fluorescence channels or the spectral overlap constraints of cyclic imaging. A single DSP run can profile up to several hundred protein targets simultaneously, with the number determined by the panel design rather than instrument optics.
ROI Selection: Fluorescence-Guided Spatial Targeting
ROI selection is the step that makes DSP spatial: you decide which tissue compartments to profile before molecular data is collected. The process works as follows:
- A tissue section is stained with a cocktail of DNA-barcoded detection antibodies (for the protein targets of interest) plus 3–4 fluorescently labeled morphology markers (commonly pan-cytokeratin for epithelium, CD45 for immune cells, and a nuclear stain such as Syto13).
- The entire section is imaged in the DSP instrument to produce a high-resolution fluorescence map.
- Using the fluorescence image as a guide, ROIs are defined — geometrically (circles, polygons), by fluorescent signal segmentation (e.g., panCK+ vs. panCK− regions), or by grid-based systematic sampling.
- A focused UV beam is directed at each selected ROI individually, cleaving the photocleavable linker on the barcoded antibodies within that region.
- Released barcodes are aspirated into a microtiter plate, one well per ROI, and amplified into an NGS library for sequencing.
This workflow decouples the imaging step (which defines where to profile) from the molecular quantification step (which measures how much of each target is present). The result is a quantitative, spatially registered protein expression dataset where each ROI represents a defined tissue compartment — tumor epithelium, immune stroma, invasive margin, or a tertiary lymphoid structure.
Protein Panels and Targets
GeoMx DSP supports modular protein panels built on a core of morphology markers plus pathway-specific or cell-type-specific modules. Core morphology markers typically include CD45, pan-cytokeratin, and a nuclear stain, supplemented by application-specific modules:
- Immune cell profiling: CD3, CD4, CD8, CD20, CD68, CD56, FOXP3, and additional immune lineage markers
- Immune activation and checkpoint targets: PD-1, PD-L1, CTLA-4, LAG-3, TIM-3, HLA-DR, Ki-67
- Tumor signaling and pathway markers: EGFR, HER2, PTEN, phosphorylated signaling proteins, apoptosis markers
- Custom targets: user-defined antibodies conjugated to DSP-compatible DNA barcodes
Panel size varies by configuration. The standard GeoMx protein assay supports dozens to over 100 targets per ROI. The GeoMx Discovery Proteome Atlas (DPA) extends this to 570+ protein targets, covering immune oncology, tumor signaling, cell death, PI3K/AKT, MAPK, and other cancer-relevant pathways.
GeoMx DSP vs. Other Spatial Platforms
The comparison below focuses on how GeoMx DSP proteomics differs from other spatial platforms that can also detect proteins, to help clarify when DSP is the right choice.
| Dimension | GeoMx DSP | CosMx SMI / Xenium | Visium |
|---|---|---|---|
| Resolution | Multi-cell ROI (100+ cells/ROI) | Single-cell / subcellular | ~55 μm spots (multi-cell) |
| Protein detection | NGS readout of barcoded antibodies; high-plex | Cyclic IF or limited protein co-detection | Not protein-native (RNA-focused) |
| Spatial targeting | User-defined ROIs based on morphology | Whole-area unbiased imaging | Unbiased capture across tissue |
| FFPE compatibility | Core strength | CosMx: validated; Xenium: FFPE compatible | Validated |
| Tissue area | Up to ~14.6 × 36.2 mm | Limited by FOV per run | 6.5 × 6.5 mm (capture area) |
| Best for | High-plex protein profiling of defined tissue compartments in FFPE cohorts | Single-cell resolution spatial profiling with targeted gene panels | Unbiased whole-transcriptome spatial discovery |
GeoMx DSP is the platform of choice when the research question demands quantitative, high-plex protein readouts from histologically defined tissue regions — particularly in FFPE cohort studies where sample number, tissue heterogeneity, and ROI-level statistical power are primary concerns. When single-cell spatial resolution or whole-transcriptome RNA profiling is the priority, CosMx SMI, Xenium, or Visium may be more appropriate; CD Genomics offers all of these platforms and advises on selection during study design.
GeoMx DSP Workflow
CD Genomics manages the complete GeoMx DSP spatial proteomics workflow from study design to data delivery.
- Study design, panel planning, and ROI strategy
Biological question, tissue type, and cohort structure are discussed. The protein panel (pre-designed modules, Discovery Proteome Atlas, or custom targets) and morphology markers (typically panCK, CD45, DNA stain, plus one application-specific marker) are selected. ROI strategy — geometric, segmented by morphology, or grid-based — is planned based on tissue architecture and hypothesis.
- Tissue sectioning and quality control
FFPE sections (4–5 μm) or fresh-frozen sections (5–10 μm) are mounted on charged slides. Section quality is assessed for tissue integrity, adhesion, and autofluorescence. Serial sections are prepared for H&E staining to guide ROI placement and for DSP staining.
- Antibody incubation and fluorescence imaging
Protein detection antibodies (each conjugated to a unique UV-cleavable DNA barcode) and fluorescent morphology markers are incubated on the tissue section in a single cocktail. The slide is imaged in the GeoMx DSP instrument at high resolution across 4 fluorescence channels, generating a composite morphology image for ROI selection.
- ROI selection and barcode release
ROIs are defined using the fluorescence image. A focused UV beam illuminates each ROI individually, cleaving the photocleavable linkers and releasing barcodes from that specific tissue region. Barcodes from each ROI are collected into separate wells of a microtiter plate.
- NGS library preparation and sequencing
Collected barcodes are PCR-amplified, purified, and sequenced on an Illumina platform. Sequencing depth is scaled to the number of ROIs and protein targets. Each sequencing read is mapped to its corresponding protein target via the barcode sequence, producing a count-based expression matrix with one row per ROI and one column per protein.
- Data processing, analysis, and reporting
Raw barcode counts are normalized, QC-filtered, and analyzed. ROI-level protein expression data is linked to spatial metadata — compartment annotation, distance from tissue boundaries, and histological features — for spatially informed statistical analysis. Results are delivered as an integrated report with publication-ready figures.
GeoMx DSP Sample Requirements
| Sample Type | Requirement | Storage & Shipping | Notes |
|---|---|---|---|
| FFPE tissue block | Standard FFPE blocks | 4°C; ship with cold packs | Core sample type for GeoMx DSP proteomics. Platform tolerates the cross-linking and degradation typical of archival FFPE. |
| FFPE tissue section | 4–5 μm; mounted on charged/plus slides | Room temperature (dried and sealed); ≤14 days recommended | Provide ≥4 serial sections per sample (1 H&E, 2 DSP, 1 backup). Avoid sections with folds, tears, or detachment. |
| Fresh-frozen tissue block | OCT-embedded; ≥3 × 3 mm cross-section | −80°C; ship on dry ice | Compatible but FFPE is the more common input for DSP protein workflows. |
| Fresh-frozen section | 5–10 μm; mounted on charged slides | −80°C; ship in slide box on dry ice | Provide ≥4 serial sections. |
Slide quality is visually inspected before processing. Sections with poor adherence, extensive folding, or tissue loss are flagged and may require re-cutting from the block.
Bioinformatics & Spatial Data Analysis
GeoMx DSP spatial proteomics data analysis moves from raw NGS-barcode counts to spatially informed biological interpretation, with every analysis step traceable by ROI.
Standard analysis
- Data preprocessing and QC: barcode count normalization, sample-level and ROI-level QC metrics (sequencing depth, barcode distribution, negative control signal), ROI filtering based on nuclei count and signal-to-noise ratio
- Dimensionality reduction and visualization: PCA, UMAP/t-SNE of ROI-level protein expression profiles, with color-coding by tissue compartment, sample, or treatment group
- Differential protein expression: linear mixed models or limma-based comparisons between ROIs grouped by compartment (tumor vs. stroma), condition (treated vs. control), or clinical variable, with multiple-testing correction
- Spatial feature mapping: protein expression values projected back onto tissue coordinates, visualizing spatial gradients and compartment-specific expression patterns
- Heatmap and pathway analysis: clustered heatmaps of differentially expressed proteins; pathway enrichment for compartment-specific protein signatures
- ROI metadata annotation: each ROI tagged with its compartment label, morphological features, and sample-level metadata for structured statistical modeling
Advanced analysis (optional)
- Multi-compartment spatial modeling: statistical testing of protein expression differences across three or more compartments within the same tissue section (e.g., tumor core vs. invasive margin vs. adjacent stroma)
- ROI-level immune cell deconvolution: inference of immune cell subtype proportions from protein marker combinations within each ROI
- Multi-omics spatial integration: correlation of GeoMx DSP protein data with spatial transcriptomics or proteomics data from serial sections for combined RNA + protein spatial profiling
- Clinical outcome association: survival analysis, response prediction, or other clinical endpoint modeling using ROI-level protein signatures
Deliverables
Every GeoMx DSP spatial proteomics project includes structured, ready-to-use deliverables.
- Processed protein expression matrix
- ROI × protein target count matrix with spatial coordinates and compartment annotations (.csv, .rds, or .h5ad)
- QC report
- Per-sample and per-ROI QC metrics, sequencing depth and barcode distribution, negative control probe signals, and ROI filtering summary
- Spatial proteomics report
- Differential protein expression results by compartment and condition, PCA/UMAP visualizations, clustered heatmaps, pathway enrichment, and spatial feature maps
- Publication-ready figures
- High-resolution spatial protein expression maps, volcano plots, multi-panel compartment comparison figures, and heatmaps (300+ dpi TIFF/PNG; editable vector PDF/SVG)
- Reproducible analysis documentation
- Fully documented R scripts with session information, enabling independent reproduction of all analyses
- Raw and intermediate data
- DSP barcode count files, fluorescence morphology images, and ROI metadata tables
GeoMx DSP Applications
Our GeoMx DSP spatial proteomics service supports a wide range of biological research questions, especially those requiring ROI-based protein profiling in FFPE tissue cohorts.
Tumor microenvironment and immuno-oncology
GeoMx DSP proteomics resolves immune cell composition and checkpoint expression across distinct tumor compartments — tumor epithelium, immune stroma, invasive margin, and tertiary lymphoid structures. This supports studies of immune exclusion mechanisms, identification of compartment-specific immunotherapy targets, and spatial biomarker validation in clinical trial cohorts. See also: Spatial Omics Solutions for Tumor Microenvironment.
Biomarker validation in FFPE cohorts
The platform's FFPE compatibility and quantitative, ROI-level protein readout make it suited for retrospective biomarker studies using archival tissue blocks. Because protein targets are measured by NGS barcode counts (not fluorescence intensity), quantification is inherently digital and batch effects can be managed through standard NGS normalization strategies. This supports multi-center cohort studies where specimens vary in age, fixation, and processing.
Immune-oncology drug development
DSP protein panels covering PD-1/PD-L1, CTLA-4, LAG-3, TIM-3, and other checkpoint targets, combined with immune lineage markers, provide a multiplexed spatial readout of the immune landscape in preclinical models and clinical specimens. ROI-level data supports pharmacodynamic analyses — for example, comparing checkpoint expression in drug-exposed vs. unexposed tissue regions within the same section.
Neuroscience and neurodegenerative disease
DSP protein panels targeting neural cell typing, Alzheimer's pathology markers (amyloid-beta, tau), and glial cell subtypes enable spatially resolved protein profiling in brain tissue sections. This supports studies of protein aggregation, neuroinflammation, and region-specific pathology in postmortem FFPE brain specimens. See also: Spatial Omics Solutions for Neuroscience.
Inflammatory and autoimmune disease
DSP resolves immune cell organization and activation states in inflamed tissues — distinguishing resident vs. infiltrating populations, identifying spatial gradients of cytokine and checkpoint expression, and mapping disease-relevant protein signatures to defined histological compartments.
Case Study: GeoMx DSP Spatial Profiling of the Immune Microenvironment in Barrett's Esophagus Progression
Source: Frei NF, Kontny U, Pohland MO, et al., Cancers, 2023
Background
The stepwise progression from Barrett's esophagus (BE) through dysplasia to esophageal adenocarcinoma (EAC) is accompanied by immune microenvironment remodeling, but the spatial organization of immune cells and signaling proteins across these disease stages is poorly characterized in FFPE archival specimens. Understanding which immune changes occur at each histological stage — and where in the tissue they localize — requires spatial protein profiling that preserves tissue architecture and compartment identity.
Methods
Endoscopic mucosal resection specimens from three patients, each containing regions of BE, dysplasia, and EAC within the same tissue section, were processed as FFPE sections for GeoMx DSP analysis. Protein expression was profiled alongside an RNA panel of 1,812 cancer-related transcripts. ROIs were selected across distinct tissue compartments — BE epithelium, dysplastic epithelium, and invasive EAC regions — guided by fluorescence morphology markers. DSP barcode counts were quantified by NGS and normalized for compartment-level protein expression comparison across disease stages.
Results
GeoMx DSP successfully quantified protein and RNA expression from all three disease stages within single FFPE sections, demonstrating the feasibility of spatial multi-omics profiling in endoscopic resection specimens. Spatial protein profiling revealed compartment-specific immune signatures that distinguished BE, dysplasia, and EAC regions — patterns that would be invisible in whole-tissue homogenate analysis. The study validated DSP as a practical tool for retrospective spatial immune profiling in esophageal pre-cancer and cancer specimens, where tissue quantity is often limited and histological heterogeneity is high.
Conclusion
This feasibility study demonstrates that GeoMx DSP can extract quantitative, spatially resolved protein expression data from small FFPE endoscopic specimens spanning the full BE–dysplasia–EAC progression. The ability to compare immune protein signatures across disease stages within the same tissue section — while preserving compartment- and stage-specific spatial context — mirrors the retrospective cohort study designs CD Genomics supports with its GeoMx DSP spatial proteomics service.
Frequently Asked Questions (FAQ)
References
- Frei NF, Kontny U, Pohland MO, et al. "Feasibility study utilizing NanoString's Digital Spatial Profiling (DSP) technology for characterizing the immune microenvironment in Barrett's esophagus formalin-fixed paraffin-embedded tissues." Cancers, vol. 15, no. 24, 2023, 5895.
- Merritt CR, Ong GT, Church SE, et al. "Multiplex digital spatial profiling of proteins and RNA in fixed tissue." Nature Biotechnology, vol. 38, 2020, pp. 586–599.
- Su DG, Schoenfeld DA, Ibrahim W, et al. "Digital spatial proteomic profiling reveals immune checkpoints as biomarkers in lymphoid aggregates and tumor microenvironment of desmoplastic melanoma." Journal for ImmunoTherapy of Cancer, vol. 12, no. 3, 2024, e008646.
- Martinez-Morilla S, Villarroel-Espindola F, Wong PF, et al. "Digital spatial profiling of melanoma shows CD95 expression in immune cells is associated with resistance to immunotherapy." OncoImmunology, vol. 12, no. 1, 2023, 2260618.
- Hernandez S, Lazcano R, Serrano A, Powell S, Kostousov L, Mehta J, Khan K, Lu W, Solis LM. "Challenges and opportunities for immunoprofiling using a spatial high-plex technology: the NanoString GeoMx Digital Spatial Profiler." Frontiers in Oncology, vol. 12, 2022, 890410.
