• HOME
  • NEWS
  • EXPLORE
    • CAREER
      • Companies
      • Jobs
    • EVENTS
    • iGEM
      • News
      • Team
    • PHOTOS
    • VIDEO
    • WIKI
  • BLOG
  • COMMUNITY
    • FACEBOOK
    • INSTAGRAM
    • TWITTER
Wednesday, November 18, 2026
BIOENGINEER.ORG
No Result
View All Result
  • Login
  • HOME
  • NEWS
  • EXPLORE
    • CAREER
      • Companies
      • Jobs
        • Lecturer
        • PhD Studentship
        • Postdoc
        • Research Assistant
    • EVENTS
    • iGEM
      • News
      • Team
    • PHOTOS
    • VIDEO
    • WIKI
  • BLOG
  • COMMUNITY
    • FACEBOOK
    • INSTAGRAM
    • TWITTER
  • HOME
  • NEWS
  • EXPLORE
    • CAREER
      • Companies
      • Jobs
        • Lecturer
        • PhD Studentship
        • Postdoc
        • Research Assistant
    • EVENTS
    • iGEM
      • News
      • Team
    • PHOTOS
    • VIDEO
    • WIKI
  • BLOG
  • COMMUNITY
    • FACEBOOK
    • INSTAGRAM
    • TWITTER
No Result
View All Result
Bioengineer.org
No Result
View All Result
Home NEWS Science News Biology

Stemness Gene Profiling Identifies Prognostic Subtypes and Drug Targets in Osteosarcoma

by
November 18, 2026
in Biology
Reading Time: 7 mins read
0
Stemness Gene Profiling Identifies Prognostic Subtypes and Drug Targets in Osteosarcoma

Stemness Gene Profiling Identifies Prognostic Subtypes and Drug Targets in Osteosarcoma

Share on FacebookShare on TwitterShare on LinkedinShare on RedditShare on Telegram

Osteosarcoma represents the most prevalent primary malignant bone tumor diagnosed in children and adolescents, exhibiting an annual worldwide incidence of roughly 4.8 cases per million individuals. Although surgical resection paired with multi-agent adjuvant chemotherapy has stabilized five-year survival rates between 60 and 70 percent for localized presentations, patients suffering from postoperative recurrence or primary metastatic dissemination face dismal outcomes, with long-term survival figures remaining stubbornly below 25 percent. A primary challenge in clinical management stems from extensive biological and molecular heterogeneity that classical histopathological classifications, such as osteoblastic, chondroblastic, and fibroblastic groupings, fail to resolve in terms of therapeutic sensitivity. Rather than being driven by recurrent actionable point mutations, the genomic framework of osteosarcoma is defined by severe chromosomal instability, chromothripsis, and complex structural rearrangements. Arising predominantly from osteogenic precursor cells along the bone marrow mesenchymal stem cell lineage, osteosarcomas frequently undergo disruptions in normal differentiation pathways, fostering the emergence of cancer stem cell populations characterized by self-renewal and multilineage potential. To dissect the clinical and biological consequences of stemness-associated transcriptional heterogeneity, a research team performed an extensive integrative investigation analyzing bulk patient cohorts alongside single-cell transcriptomes, establishing distinct molecular subtypes and developing a validated prognostic gene expression signature.

To rigorously evaluate the transcriptional architecture of stemness across independent clinical cohorts, the investigators compiled and standardized genomic datasets from the public TARGET-OS and GSE21257 repositories, comprising an integrated cohort of 138 osteosarcoma patients with available overall survival data. The raw RNA sequencing read counts from TARGET-OS, encompassing 85 primary tumor samples, underwent variance-stabilizing transformation using DESeq2, whereas probe-level expression measurements from the 53 microarray profiles in GSE21257 were mapped directly to gene symbols. To overcome batch effects across different sequencing and microarray platforms, the investigators retained 13,768 intersecting genes and applied the ComBat algorithm without incorporating survival outcomes into the adjustment model. To capture the multidimensional nature of stemness, the team aggregated 1,527 stemness-associated features from the StemChecker platform, encompassing mesenchymal stem cell signatures, embryonic stem cell core transcription factor target sets for NANOG, OCT4, and SOX2, the PluriNet regulatory network, and curated developmental pathways from KEGG and Reactome. Utilizing non-negative matrix factorization across 50 repeated runs with the Brunet algorithm, factorization quality metrics across candidate ranks from two to ten identified an optimal two-cluster partition, classifying patients into two exploratory stemness subtypes designated Cluster 1 and Cluster 2.

The two stemness-associated subtypes revealed striking divergences in their anatomical distributions and clinical outcomes within the TARGET-OS cohort. Notably, all patients categorized into Cluster 1 presented exclusively with primary tumors localized to long bones, whereas 14.3 percent of patients in Cluster 2 presented with primary tumors originating at non-long bone sites, including 9.5 percent in the pelvis and 4.8 percent across other anatomical regions. Survival analyses demonstrated that Cluster 1 experienced significantly inferior overall survival compared to Cluster 2. In a complete-case multivariable Cox proportional hazards analysis of 84 patients recording 29 deaths, stratified by metastatic status at diagnosis to satisfy proportional hazards assumptions and adjusted for primary tumor site, Cluster 1 retained a significant association with poorer overall survival, displaying a hazard ratio of 2.71 with a 95 percent confidence interval spanning 1.13 to 6.51. When further adjusted for patient age as a continuous variable and biological sex, the adverse prognostic impact of Cluster 1 remained pronounced, yielding a hazard ratio of 2.83 with a 95 percent confidence interval of 1.16 to 6.88, demonstrating that stemness classification provides independent prognostic value beyond standard clinical indicators.

Differential expression and functional pathway characterization between the subtypes uncovered distinct molecular programs driving the observed clinical divergence. Differential expression analysis in DESeq2 identified 2,177 significantly upregulated genes in Cluster 1 and 1,481 significantly upregulated genes in Cluster 2 under thresholds of an adjusted p-value below 0.05 and an absolute log2 fold-change exceeding 1. Reactome pathway enrichment demonstrated that genes elevated in Cluster 1 were enriched for membrane receptor signaling, ion transport cascades, and lipid metabolism pathways, while genes upregulated in Cluster 2 corresponded to extracellular matrix remodeling, complement cascades, and calcium signaling. Complementary gene set variation analysis using the Hallmark collection confirmed that Cluster 1 exhibited heightened activity across metabolic and genomic maintenance pathways, including oxidative phosphorylation, fatty acid metabolism, DNA repair machinery, and MYC targets. Conversely, Cluster 2 demonstrated enriched activation of epithelial-mesenchymal transition, hedgehog signaling, and ultraviolet response pathways, indicating that Cluster 1 is governed by hyperactive proliferative and energetic signaling whereas Cluster 2 is characterized by extensive matrix remodeling and immune microenvironment engagement.

An evaluation of somatic genomic landscapes and computational drug response predictions further differentiated the two stemness subtypes. Analysis of somatic single-nucleotide variants across the TARGET-OS cohort revealed that missense mutations predominated, accompanied by nonsense, frameshift, and splice-site alterations, with cytosine-to-thymine transitions forming the primary mutational signature. While TP53, ATRX, RB1, TTN, and MUC16 emerged as the most frequently altered genes overall, subtype-specific comparisons indicated that TP53 and MUC16 mutations were numerically more frequent in Cluster 2, whereas RB1 and ATRX mutations appeared more frequently in Cluster 1, though none reached statistical significance after genome-wide Benjamini-Hochberg adjustment. The overall tumor mutational burden remained comparable between the subtypes, showing a median of 0.42 mutations per megabase in Cluster 1 versus 0.34 mutations per megabase in Cluster 2. Within Cluster 1, the presence of somatic RB1 mutations was associated with worsened survival, showing a hazard ratio of approximately 4.16. Phenotypic drug screening models generated via the oncoPredict package utilizing pre-trained CTRP2 data predicted that Cluster 1 was uniquely sensitive to bosutinib and fixed-ratio combinations containing the MEK inhibitor selumetinib combined with decitabine, tretinoin, or piperlongumine. In contrast, Cluster 2 demonstrated heightened predicted sensitivity to multiple receptor tyrosine kinase inhibitors including cabozantinib, axitinib, pazopanib, and lenvatinib, corroborating distinct actionable signaling dependencies.

To explore the broader systems-level transcriptional architecture supporting the stemness subtypes, the authors conducted weighted gene co-expression network analysis across the top 15,000 variable genes in the TARGET-OS cohort. Setting a soft-thresholding power of 3 to achieve an approximate scale-free topology fit index of 0.85, hierarchical clustering partitioned the co-expression landscape into 11 distinct co-expression modules. Module-trait correlation analyses revealed that the brown and red modules correlated positively with the aggressive Cluster 1 subtype, whereas the yellow module exhibited a robust positive correlation with Cluster 2. Protein-protein interaction networks constructed via the STRING database combined with Reactome enrichment demonstrated that core hub genes within the brown and red modules mediated membrane receptor signaling, transmembrane ion transport, and the release of inflammatory mediators. In contrast, hub genes within the yellow module were enriched in extracellular matrix organization, focal adhesion, cellular migration, and survival pathways. Subsequent extraction of subtype-specific differentially expressed genes directly mapped to these functional modules isolated 60 core genes tightly linked to Cluster 1 biology and 33 genes defining the microenvironmental features of Cluster 2.

Building upon the biological insights gained from stemness subtyping, the investigators constructed a quantitative machine learning prognostic signature to predict overall survival. Beginning with 1,400 stemness-related candidate genes present within the merged TARGET-OS and GSE21257 matrix, univariate Cox regression in the TARGET-OS training set identified 55 genes significantly associated with overall survival at a significance threshold of p below 0.01, with hazard ratios spanning from 0.20 to 4.20. Following proportional hazards assumption testing, five genes exhibiting non-proportional hazards were excluded, leaving 50 candidate features for LASSO-penalized Cox regression. Cross-validation tuning isolated 22 non-zero candidate genes, which were subsequently refined via bidirectional stepwise Cox selection based on the Akaike information criterion to establish a parsimonious nine-gene signature. Within this final model, PIP5K1C, SNX1, TMEM43, and ZNF300 carried negative coefficients and acted as protective factors involved in phospholipid signaling, endocytosis, and transcriptional homeostasis. Conversely, STC2, SERPINH1, RAD23A, TRPS1, and SASH1 carried positive coefficients and functioned as risk factors implicated in cellular stress adaptation, proteostasis, DNA repair, and invasive migration.

The quantitative risk score derived from the nine-gene signature demonstrated robust performance in risk-stratifying osteosarcoma patients across independent cohorts. Applying the median risk score from the TARGET-OS training cohort as an absolute cutoff, high-risk patients exhibited significantly reduced overall survival compared to low-risk counterparts. In the training cohort, time-dependent receiver operating characteristic curves demonstrated area under the curve values of 0.917 at one year, 0.923 at three years, and 0.927 at five years. When applied to the external GSE21257 validation cohort without recalibrating model coefficients, the signature maintained discriminatory capacity, achieving area under the curve values of 0.653, 0.751, and 0.711 at one, three, and five years, respectively. Examination of risk score distributions demonstrated that high-risk tumors were heavily concentrated within the Cluster 1 subtype, whereas low-risk tumors predominated in Cluster 2. Furthermore, combining the nine-gene risk score with clinical variables including metastatic presentation into a predictive nomogram achieved an area under the curve of 0.988 for one-year survival in the training cohort, markedly outperforming the risk score alone at 0.900 and metastatic status alone at 0.925.

Immunogenomic deconvolution across multiple algorithmic platforms revealed pronounced disparities in the tumor immune microenvironment between high- and low-risk patients. Implementation of the ESTIMATE algorithm established that high-risk osteosarcomas had significantly lower overall immune scores, reduced stromal scores, and elevated tumor purity relative to low-risk specimens. Detailed cellular deconvolution employing MCP-counter, quanTIseq, EPIC, and CIBERSORT demonstrated consistent reductions in estimated CD8+ T Cell and B Cell infiltration within high-risk tumors, accompanied by lower MCP-counter cytotoxic lymphocyte scores. Concurrently, differential gene expression showed that low-risk tumors were enriched for functional pathways governing adaptive humoral immunity, lymphocyte receptor signaling, and somatic recombination, with significant upregulation of the immune checkpoint molecules SIRPA, BTN2A2, and CD27. In contrast, high-risk tumors exhibited marked downregulation of the cytotoxic effectors PRF1 and GZMB, the T-cell markers CD8A and CD8B, and antigen presentation molecules including HLA-A and B2M, indicating that poor survival in the high-risk group is intimately connected to an immune-excluded microenvironment deficient in antigen presentation and cytotoxic lymphocyte activity.

To pinpoint the cellular origin of stemness signatures at single-cell resolution, the authors interrogated scRNA-seq datasets GSE152048 and GSE162454 encompassing 127,196 rigorously filtered cells spanning ten lineages. Calculation of the nine-gene signature across individual cells revealed that risk scores were highest in malignant osteosarcoma cells and cancer-associated fibroblasts, whereas tumor-associated macrophages and lymphocytes exhibited lower scores. Subclustering of malignant osteosarcoma cells identified five distinct subpopulations, designated OS1 through OS5, with the OS1 subpopulation carrying the highest signature scores and positioning at the terminus of inferred differentiation trajectories. Functional enrichment of OS1-upregulated genes highlighted pathways governing ossification and extracellular matrix production. Furthermore, CellChat ligand-receptor communication modeling predicted robust interactions between malignant osteosarcoma cells, cancer-associated fibroblasts, and monocytes via the COL1A1-CD44 signaling axis. The authors noted that retrospective cohort sizes, a drop in performance between training and validation cohorts, and the reliance on shared osteogenic markers represent notable study limitations, emphasizing that these computational findings, cellular interactions, and subtype-specific drug sensitivities warrant dedicated prospective experimental validation.

Subject of Research: Stemness-associated transcriptional profiling, prognostic modeling, and microenvironmental characterization in osteosarcoma

Article Title: Stemness-associated transcriptional profiles reveal prognostic and microenvironmental heterogeneity in osteosarcoma

Article References: Cui, L., Liu, X., Yu, F., Zhu, Y., Lian, Q., Wang, H., Fu, G., & Zheng, X. (2026). Stemness-associated transcriptional profiles reveal prognostic and microenvironmental heterogeneity in osteosarcoma. Frontiers in Genetics, 17, Article 2027108. https://doi.org/10.3389/fgene.2026.2027108

Image Credits: AI Generated

DOI: 10.3389/fgene.2026.2027108

Keywords: Osteosarcoma, Cancer Stemness, Transcriptomics, Prognostic Signature, Tumor Microenvironment, TARGET-OS, GSE21257, Non-Negative Matrix Factorization, Single-Cell RNA Sequencing, Immune Infiltration, LASSO Regression, Drug Sensitivity Prediction, Cancer-Associated Fibroblasts, COL1A1-CD44, Cell-Cell Communication, Pediatric Bone Cancer, Chemoresistance

News Source: Juliet Wilcox. (November 18, 2026). Stemness Gene Profiling Identifies Prognostic Subtypes and Drug Targets in Osteosarcoma. Scienmag.

Tags: Cancer StemnessCancer-associated fibroblastscell-cell communicationchemoresistanceCOL1A1-CD44drug sensitivity predictionGSE21257immune infiltrationLASSO regressionnon-negative matrix factorizationOsteosarcomapediatric bone cancerprognostic signaturesingle-cell RNA sequencingTARGET-OSTranscriptomicstumor microenvironment
Share12Tweet7Share2ShareShareShare1

Related Posts

New Flying Frog Species Discovered in Indonesia's Mekongga Mountains

New Flying Frog Species Discovered in Indonesia’s Mekongga Mountains

November 18, 2026
Adaptive Pseudo-References Improve Copy Number Variant Detection in Low-Input Sequencing

Adaptive Pseudo-References Improve Copy Number Variant Detection in Low-Input Sequencing

November 18, 2026

TopoPPI Reduces Residue Footprint Fragmentation Across Complex Protein Interfaces

November 18, 2026

African Simian Arteriviruses Pose Potential Cross-Species Transmission Risk to Humans

November 18, 2026

POPULAR NEWS

  • Multi-Environment Trials Identify Four Superior High-Yielding Barley Genotypes in Egypt

    Multi-Environment Trials Identify Four Superior High-Yielding Barley Genotypes in Egypt

    29 shares
    Share 12 Tweet 7
  • Shandong Listeria Study Links Food and Clinical Isolates Using Genomic Profiling

    29 shares
    Share 12 Tweet 7
  • Engineers Develop Curved Lithium-Sulfur Battery Electrodes for Drone Frame Energy Storage

    29 shares
    Share 12 Tweet 7
  • Maternal Serum Elements Link to Cardiometabolic Risk Factors in Pregnancy

    29 shares
    Share 12 Tweet 7

About

We bring you the latest biotechnology news from best research centers and universities around the world. Check our website.

Follow us

Recent News

Armenian Medical Education Transformation During Concurrent Pandemic and Conflict Crises

New Flying Frog Species Discovered in Indonesia’s Mekongga Mountains

Transcription Factor Networks Coordinate Molecular Signaling for Pepper Cold Stress Resilience

Subscribe to Blog via Email

Enter your email address to subscribe to this blog and receive notifications of new posts by email.

Join 88 other subscribers
  • Contact Us

Bioengineer.org © Copyright 2023 All Rights Reserved.

Welcome Back!

Login to your account below

Forgotten Password?

Retrieve your password

Please enter your username or email address to reset your password.

Log In
No Result
View All Result
  • Homepages
    • Home Page 1
    • Home Page 2
  • News
  • National
  • Business
  • Health
  • Lifestyle
  • Science

Bioengineer.org © Copyright 2023 All Rights Reserved.