Cancer ResearchCancer BiologyPeer-reviewed journal article
Spatial profiling identifies regionally distinct microenvironments and targetable immunosuppressive mechanisms in pediatric osteosarcoma pulmonary metastases.
Jason EigenbroodiD, Nathan W. WongiD, Paul W. MalloryiD, Janice S. PereiraiD, Demond WilliamsiD, Douglass W. Morris-IIiD, Jessica A. BeckiD, James C. CronkiD, Carly M. SayersiD, Monica MendeziD, Linus KaiseriD, Julie GalindoiD, Jatinder SinghiD, Ashley CardamoneiD, Milind M. PoreiD, Michael C. KellyiD, Amy Kathleen LeBlanciD, Jennifer A. CotteriD, Rosandra Natasha KaplaniD, Troy A. McEachroniD
In-depth summary
Claude Opus 5.5 · xhigh effort · 8 Nov 2026 · from the full text of version 1 (24 Jan 2025) · CC0 1.0
Visium profiling of nine pediatric osteosarcoma pulmonary metastases reveals a conserved granuloma-like architecture: inflamed macrophage-rich periphery, collagen capsule, inflammation-poor core, and CXCR4 as candidate target.
Metastatic osteosarcoma has poor outcomes, yet beyond well-catalogued genomic alterations little is known about how its metastatic microenvironment is spatially organized or how it suppresses immunity, which single-cell studies of mostly primary tumors cannot resolve. This work maps region-specific transcriptional programs in archived FFPE lung metastases and uses them to nominate spatially targeted therapies.
Key results
ShownDespite the disease's reputation for genomic heterogeneity, the metastases share a reproducible spatial organization (intratumoral core clusters, a circumscribing peritumoral zone, other extratumoral tissue) whose region-specific programs persist after cross-specimen integration.
Nine FFPE specimens from seven patients (seven lung, two hilar lymph node); per-specimen Leiden clusters split into intratumoral vs extratumoral groups, with a peritumoral ring in all but fibroblastic OS-7; Harmony integration gave 10 integrated clusters, with immune/inflammatory pathways extratumoral and cell cycle/ossification intratumoral. · Resolving the spatial transcriptome of…; Data integration identifies a conserved…; Figure legends
SuggestedLesions in both lung parenchyma and hilar lymph nodes are wrapped by a foreign body granuloma-like response: giant cells, a collagen capsule, a peripheral defense-response program, and concentric macrophage, myofibroblast and CD8 T-cell layers, with MARCO+ macrophages confined to the margin.
Langhans-like multinucleated giant cells on H&E; trichrome collagen encircling the core; MSigDB Defense Response signature enriched outside the core and inversely correlated with the 23-gene Osteosarcoma Signature (Spearman, significance p ≤ 0.05); IMC showed CD68+, SMA+ and CD8+ cells concentric at the margin; MARCO+ macrophages peritumoral, absent intratumorally by IHC. · Evidence of a peritumoral foreign body-like…; MARCO expressing macrophages colocalize with the…; Figure legends
ShownThe tumor core is an inflammatory desert: inferred IFNγ and TNFα activity and inflammatory ligand-receptor traffic sit outside it, nuclear p65 is absent within it, and tumor cells express few inflammatory chemokines and cytokines, while growth factor activity localizes intratumorally.
CytoSig correlations split each specimen into intratumoral and extratumoral cluster groups; IFNγ/TNFα activity was extratumoral only; 105-analyte arrays on independent tumor lysates and patient-derived line media left most inflammatory analytes undetectable (SPP1, VEGFA, MIF, EMMPRIN high); no nuclear p65 in tumor; CellChat ranked peritumoral clusters most communicative. · Regionally restricted chemokine, cytokine, and…; The intratumoral microenvironment of metastatic…; Inflammatory signaling is enriched in the…
ShownTGFβ-family signaling is pervasive but heterogeneous across regions, with BMP6 and BMP7 activity favoring the tumor core and canonical SMAD4-dependent signaling active intratumorally.
CytoSig inferred TGFβ activity throughout all tissue regions while BMP6 and BMP7 activity was largely intratumoral; CellChat recovered a complex TGFβ-family interactome in each specimen; IHC detected nuclear SMAD4 within the intratumoral region of all specimens. · Widespread and heterogeneous signaling activity…; Figure legends
SuggestedSpatial drug-gene enrichment nominates the CXCR4 antagonist plerixafor, chiefly against a peritumoral myeloid-suppressive cluster, while peripheral CXCL12 and intratumoral MIF expression plus ligand-receptor analysis suggest regionally divergent canonical and non-canonical CXCR4 signaling.
IDG Drug Targets 2022 (EnrichR) queried with integrated-cluster markers ranked plerixafor highest, mainly for IC-3 (weaker for IC-5, IC-6; IC-3 and IC-6 form the peritumoral zone); CXCL12 RNA peripheral vs MIF intratumoral; MIF protein detected, CXCL12 not; CXCR4 moderately linked to worse overall survival in primary-tumor data. · Drug-gene interaction analysis identifies CXCR4…; Regionally distinct expression patterns of…; The intratumoral microenvironment of metastatic…
Methods, data and reproducibility
Visium CytAssist whole-transcriptome profiling of archived FFPE metastases, with per-specimen SCTransform/Leiden clustering, a derived 23-gene intratumoral signature, UCell signature scoring, CytoSig cytokine-activity inference, CellChat ligand-receptor analysis, Harmony integration and drug-gene enrichment, validated by IMC, IHC, trichrome and cytokine arrays.
- FFPE osteosarcoma pulmonary metastases (lung parenchyma and hilar lymph nodes) from pediatric patients at Children's Hospital Los Angeles, profiled by Visium CytAssist · human · n = 9 specimens from seven patients
- Independent cohort of frozen osteosarcoma pulmonary metastasis tissue homogenates (high tumor cellularity) profiled on Human XL Cytokine Arrays covering 105 analytes · human
- Patient-derived osteosarcoma cell lines; conditioned media from three biological replicates pooled per line for cytokine arrays · human
- Clinomics 'RNAseq_Landscape_Manuscript' project (NCI Oncogenomics Section), primary-tumor expression with outcome, used for CXCR4 Kaplan-Meier analysis · human
- Sorenson et al. NanoString IO360 dataset comparing chemokine/cytokine gene expression in non-metastatic vs pulmonary metastatic osteosarcoma · human
- Human Lung Cell Atlas and Human PBMC reference signatures (via Azimuth) used for UCell cell-type scoring and signature curation · human
- Methods & tools
- Visium CytAssist Spatial Gene Expression (10x Genomics), Human Transcriptome Probe Set v2, SpotClean, Seurat (SCTransform, Leiden clustering, FindAllMarkers), STDeconvolve, Escape and UCell (MSigDB Hallmarks, Reactome, Biological Processes), CytoSig, CellChat, Harmony, EnrichR (IDG Drug Targets 2022, KEGG 2021, Reactome 2022, WikiPathways 2023), Imaging mass cytometry (Hyperion) with IMC-Denoise and Bodenmiller IMC Segmentation Pipeline, Human XL Cytokine Array (R&D Systems), IHC (SMAD4, p65, MARCO), trichrome staining and RNAscope GAPDH quality control
- Reagents & environment
- R v4.3.0; Seurat v4.2 to v5.1; SpaceRanger v2.0.0 (Bcl2fastq 2.20.0, STAR2.7.2a); RTA 3.4.4; refdata-gex-GRCh38-2020-A with Visium Human Transcriptome Probe Set v2.0; Prism 10.
- Reproducibility gaps
- GEO deposition of the Visium expression data was still in progress at posting, so no accession is given; no custom analysis code repository is provided, only published third-party tools and web servers are cited.
Limitations (6)
- Small cohort (nine specimens), and the 55µm Visium spot size precludes single-cell resolution, so intratumoral tumor and stromal subsets with overlapping gene signatures could not be characterized individually. · stated by the authors
- The CXCR4 survival association was assessed on tumor tissue from primary sites rather than from metastatic lesions. · stated by the authors
- Whether the foreign body-like response is specific to lung and lymph node sites, and what the capsule does functionally, remain open; more anatomical sites, in vivo experiments and single-cell spatial profiling are needed. · stated by the authors
- The functional impact of CXCR4 signaling and its suitability as a therapeutic target were not tested experimentally; the authors call for genetically engineered or syngeneic murine models. · stated by the authors
- Inferred cytokine activity was not always confirmed at the protein level (IL12 p70 inferred intratumorally but undetectable by array), and arrays used bulk lysates from a separate cohort, so CytoSig-based regional claims partly rest on unvalidated inference. · noted in this summary
- Inclusion required ample adjacent non-malignant tissue and excluded necrotic, hemorrhagic or RNA-poor blocks, which may favor well-demarcated lesions and inflate the apparent cross-specimen conservation. · noted in this summary
Read this if you study osteosarcoma metastasis or tumor immune exclusion, or want an example of Visium CytAssist on archived FFPE pediatric sarcoma tissue linked to drug-target nomination.
- Citations
- 21
- FWCI
- 13.88
- Field percentile
- top 1%
Abstract
Osteosarcoma is the most common malignant bone tumor in young patients and remains a significant clinical challenge, particularly in the context of metastatic disease. Despite extensive documentation of genomic alterations in osteosarcoma, studies detailing the immunosuppressive mechanisms within the metastatic osteosarcoma microenvironment are lacking. Our objective was to characterize the spatial transcriptional landscape of metastatic osteosarcoma to reveal these immunosuppressive mechanisms and identify promising therapeutic targets. Here, we performed spatial transcriptional profiling on a cohort of osteosarcoma pulmonary metastases from pediatric patients. We reveal a conserved spatial gene expression pattern resembling a foreign body granuloma, characterized by peripheral inflammatory signaling, fibrocollagenous encapsulation, lymphocyte exclusion, and peritumoral macrophage accumulation. We also show that the intratumoral microenvironment of these lesions lack inflammatory signaling. Additionally, we identified CXCR4 as an actionable immunomodulatory target that bridges both the intratumoral and extratumoral microenvironments and highlights the spatial heterogeneity and complexity of this pathway. Collectively, this study reveals that metastatic osteosarcoma specimens are comprised of multiple regionally distinct immunosuppressive microenvironments.
- organism: human
- assay: spatial transcriptomics
- assay: imaging
- research focus: spatial domain detection
- research focus: cell-cell communication
- research focus: clustering
- research focus: drug response / drug discovery
- approach: experimental assay
Reproducibility ledger
Nobody has reported trying to reproduce this paper yet. Did you replicate an experiment or run the code? Your report saves the next lab a week.
Sign in to file a reproducibility report.
Related papers
- Immunosuppressive tumor microenvironment of osteosarcoma
A. M. Taylor, J. Sheng, … C. C. Lau·bioRxiv·Nov 1, 2023
- Metastasis-initiating osteosarcoma subpopulations establish paracrine interactions with both lung and tumor cells to create a metastatic niche
James Brandon Reinecke, Leyre Jimenez Garcia, … Ryan David Roberts·Cancer ResearchPeer-reviewed·Aug 29, 2025
- Single-Cell and Spatial Profiling of Tumor Microenvironment Heterogeneity in Human Osteosarcomas
Xuejing Zheng, Jinbo Yuan, … Wence Wu·bioRxiv·Jun 27, 2025
- Integrative Single-cell and Spatial Transcriptomic Analysis of Osteosarcoma Reveals Conserved and Distinct Ecosystems Across Sites and Species
Yogesh Budhathoki, Matthew V. Cannon, … Ryan David Roberts·bioRxiv·May 8, 2026
- Aberrant activation of wound healing programs within the metastatic niche facilitates lung colonization by osteosarcoma cells
J. B. Reinecke, L. Jimenez Garcia, … R. D. Roberts·bioRxiv·Jan 10, 2024
- The Dual Roles of the CXCL10-CXCR3 Axis and Its Therapeutic Potential in Osteosarcoma
Benjamin Baffour Gyau, Junyan Wang, … Tsz‐Kwong Man·Journal of bone oncologyPeer-reviewed·May 23, 2025