Proteogenomic profiling of soft tissue leiomyosarcoma reveals distinct molecular subtypes with divergent outcomes and therapeutic vulnerabilities

This study establishes a proteogenomic framework for soft tissue leiomyosarcoma by identifying three distinct molecular subtypes with unique signaling pathways, clinical outcomes, and therapeutic vulnerabilities, including a validated immunohistochemical classifier for clinical application.

The Gist

Imagine Soft Tissue Leiomyosarcoma (STLMS) not as a single, uniform monster, but as a chaotic city with three very different neighborhoods. For years, doctors treated all patients in this city the same way, but many still got sick or died. This paper is like a high-tech detective squad that finally mapped out these three neighborhoods, revealing that they are built on completely different blueprints, run by different gangs, and require different strategies to defeat.

Here is the story of their discovery, broken down into simple concepts:

1. The Three Neighborhoods (The Subtypes)

The researchers took a deep look at 72 tumor samples using a "proteogenomic" microscope. This means they didn't just look at the DNA (the blueprint); they looked at the proteins (the workers) and their active switches (phosphorylation) to see what the tumor was actually doing in real-time. They found three distinct types of tumors, which they named P1, P2, and P3.

• Neighborhood P1 (The Quiet Suburb):

  • The Vibe: This is the "good" neighborhood. The cells are relatively calm, not growing too fast, and their DNA is stable (not falling apart).
  • The Strategy: They rely on specific fuel sources (metabolic pathways) to survive.
  • The Outcome: Patients with this type tend to do much better. It's the easiest to manage.

• Neighborhood P2 (The Chaotic War Zone):

  • The Vibe: This is the most dangerous neighborhood. It's a riot. The cells are growing wildly, the DNA is a mess, and the neighborhood is full of inflammatory noise (immune signals that are actually confusing the body).
  • The Strategy: They are running on high-octane engines (RTK-RAS pathways) and have a "suppressive shield" (immune evasion) that tricks the body's police force (immune system) into thinking everything is fine, even while they recruit "bad cop" macrophages to help them hide.
  • The Outcome: This is the worst type. Patients here have the shortest survival times because the tumor is aggressive and hides well.

• Neighborhood P3 (The Construction Site):

  • The Vibe: This neighborhood is under constant, frantic construction. The cells are obsessed with building new copies of themselves (cell cycle) and fixing broken DNA.
  • The Strategy: They are constantly repairing their own broken DNA using a "backup plan" (Non-Homologous End Joining) because their main repair crew is missing. They are also very good at hiding from the immune system by turning down the lights (low immune score).
  • The Outcome: Like P2, this is a bad neighborhood with poor survival rates, but for different reasons (too much growth and DNA chaos).

2. The "DNA Damage" Mystery

Imagine DNA as a long instruction manual. Sometimes, pages get torn out (mutations) or the whole book gets duplicated by mistake (chromosomal instability).

• P1 has a mostly intact manual.
• P2 and P3 have manuals that are shredded and photocopied randomly.
• The Twist: Because P2 and P3 are so damaged, they are desperate to fix their DNA. They use a "duct tape" repair method (Non-Homologous End Joining) instead of the precise "tailor" method. This makes them vulnerable. If you cut the duct tape (using drugs like PARP inhibitors), the whole manual falls apart, and the tumor dies. This is a potential new treatment strategy.

3. The "Bad Cops" (Immune Evasion)

In the P2 neighborhood, the tumor is actually "loud" (immune-hot), meaning the body's immune system is trying to attack. However, the tumor is smart. It puts up a "Do Not Disturb" sign (a protein called LGALS9) and hires "bad cops" (M2-like macrophages) to stand guard and tell the immune system to stand down. It's a trap: the immune system is there, but it's been bribed to do nothing.

4. The New "ID Card" System

The biggest breakthrough isn't just knowing these neighborhoods exist; it's that the researchers built a simple ID card system (a test using standard pathology slides) to tell them apart.

• Before: You needed a super-expensive, complex lab test to know which neighborhood a patient lived in.
• Now: They developed a checklist of 6 simple markers (like checking for specific proteins on the cell surface) that can be done in any regular hospital lab.
• The Result: If a patient has this "ID card," doctors can immediately predict if they are in the "Quiet Suburb" (P1) or the "War Zone" (P2/P3) and choose the right treatment.

The Bottom Line

This paper is a game-changer because it stops treating all soft tissue sarcomas as the same enemy.

• P1 patients might just need standard surgery and observation.
• P2 patients might need drugs that target their specific "fuel lines" or break their "immune shields."
• P3 patients might benefit from drugs that exploit their desperate need to fix their broken DNA.

By understanding the unique "personality" of each tumor, doctors can finally move from a "one-size-fits-all" approach to a precision medicine strategy, giving patients the specific key needed to unlock their specific type of cancer.

Technical Summary

1. Problem Statement

Soft tissue leiomyosarcoma (STLMS) is an aggressive malignancy with poor prognosis, particularly in advanced or metastatic stages. Current clinical management relies heavily on anatomical site and histological grade, lacking robust molecular subclassification. Previous genomic studies have identified alterations in pathways like PI3K/AKT/mTOR, but therapeutic efficacy remains limited due to indirect pathway upregulation and tumor heterogeneity. Furthermore, existing proteomic studies have been constrained by:

• Use of formalin-fixed, paraffin-embedded (FFPE) tissues, which degrade phosphoprotein signals.
• Inclusion of uterine leiomyosarcoma, obscuring the specific biology of extra-uterine (soft tissue) LMS.
• Lack of integrated multi-omic data (DNA, RNA, total protein, and phosphoprotein) specifically for STLMS.
• Limited analysis of metastatic progression and paired primary-metastatic samples.

2. Methodology

The study employed an integrative proteogenomic approach across two independent cohorts and external validation datasets:

• Cohort 1 (Discovery): 72 fresh-frozen STLMS samples (primary and metastatic), including 18 paired primary-metastatic lesions.
  • Multi-omics: Global proteomics and phosphoproteomics via LC-MS/MS (TMT 16-plex labeling); Targeted DNA sequencing (MSK-IMPACT) for 69 samples; RNA-seq for a subset (n=27).
  • Analysis: Non-negative matrix factorization (NMF) for subtype clustering; Kinase activity inference (RoKAI); Homologous Recombination Deficiency (HRD) scoring; Immune profiling (ESTIMATE, IHC).
• Cohort 2 (Validation): 219 FFPE STLMS cases assembled into Tissue Microarrays (TMA).
  • Analysis: Immunohistochemistry (IHC) to validate subtype-specific markers and develop a clinical classifier.
• External Validation: TCGA STLMS cohort (n=53) for transcriptomic comparison and survival analysis.
• Computational Framework: Integrated analysis of copy number alterations (CNA), mRNA, protein abundance, and phosphorylation states to define regulatory networks and kinase dependencies.

3. Key Contributions

• Definition of Three Proteomic Subtypes: Identified three biologically distinct subtypes (P1, P2, P3) based on integrated proteome and phosphoproteome data, which correlate with clinical outcomes.
• Phosphoproteomic Insight: Uncovered kinase signaling networks and pathway activities that are not predictable from genomic or transcriptomic data alone.
• Metastatic Evolution: Characterized proteogenomic changes between paired primary and metastatic lesions, revealing increased genomic instability and cell cycle activity in metastases.
• Clinical Translation: Developed and validated a 6-marker IHC-based classifier (CD133, MYC, CDK2, AURKA, AURKB, HLA-ABC) capable of assigning proteome-defined subtypes in standard FFPE pathology samples.

4. Key Results

A. Molecular Subtypes and Clinical Outcomes

• P1 (Stable/Metabolic): Characterized by genomic stability, low proliferation, and enrichment of FGFR2 and PDK-associated signaling. Associated with the best recurrence-free survival (RFS) and overall survival (OS).
• P2 (Inflammatory/RTK-RAS): Genomically unstable with high chromosomal instability (CIN). Enriched for inflammatory responses, NF-κB, EMT, and RTK-RAS pathways (IGF1R, PDGFRA alterations). Shows the poorest clinical outcomes. Features an "immune-hot" but suppressive microenvironment (high LGALS9, M2-like macrophages).
• P3 (Proliferative/Repair): Genomically unstable with the highest frequency of arm-level deletions. Dominated by cell cycle activation, DNA repair programs, and smooth muscle differentiation markers. Also associated with poor outcomes.

B. Genomic and Epigenetic Landscape

• Instability: P2 and P3 exhibit higher ploidy, whole-genome duplication (WGD), and CIN compared to P1.
• Driver Alterations: TP53 (68%) and RB1 (41%) are frequently altered. NCOR1 (35%) and MAP2K4 (33%) alterations are frequent, particularly in P2/P3.
• NCOR1: High NCOR1 expression correlates with stemness markers (CD133/PROM1), smooth muscle differentiation, and YAP1/TEAD signaling.

C. DNA Repair and HRD

• HRD Status: P2 and P3 are "HRD-high" (Homologous Recombination Deficiency), while P1 is "HRD-low."
• Mechanism: P2/P3 show frequent Loss of Heterozygosity (LOH) in HRR genes (e.g., BRCA2, RAD51).
• Compensatory Pathways: HRD-high tumors (especially P3) show upregulation of Non-Homologous End Joining (NHEJ) components (XRCC5/6, PARP1).
• Metastasis: Paired metastatic lesions in P3 showed significantly higher HRD scores than matched primary lesions.
• Prognosis: Low expression of RAD51 and BRCA2 (indicating HRD) is significantly associated with poorer RFS.

D. Immune Microenvironment

• P3/Metastasis: "Immune-cold" phenotype with downregulated antigen-presenting machinery (APM) and lower ImmuneScores.
• P2: "Immune-hot" but suppressive. High infiltration of M2-like macrophages (CD163+) and elevated expression of the immunosuppressive ligand LGALS9.

E. Therapeutic Vulnerabilities

• P1: Potential sensitivity to FGFR2 inhibitors or metabolic/PDK pathway inhibitors.
• P2: Vulnerable to RTK-RAS pathway inhibitors (IGF1R/PDGFRA), CDK/Aurora kinase inhibitors, and mTOR/ERK pathway blockers.
• P3: Vulnerable to CDK/Aurora kinase inhibitors and potentially PARP inhibitors due to HRD and NHEJ reliance.

5. Significance

• Precision Medicine Framework: This study establishes the first proteogenomic framework for STLMS, moving beyond histology to define molecular subtypes with distinct therapeutic vulnerabilities.
• Biomarker Development: The validated 6-marker IHC classifier provides a practical, cost-effective tool for clinical stratification in routine pathology, bridging the gap between complex proteogenomic discovery and clinical application.
• Therapeutic Strategy: The findings nominate specific targets (CDKs, AURKA/B, PARP1, IGF1R, LGALS9) for future clinical trials, suggesting that "one-size-fits-all" approaches are insufficient for STLMS.
• Metastatic Insight: The identification of increased HRD and cell cycle activity in metastatic lesions suggests that metastatic progression is driven by specific genomic and proteomic shifts, offering new targets for preventing or treating metastasis.

In conclusion, this research redefines STLMS heterogeneity through a multi-omic lens, identifying three distinct subtypes with divergent prognoses and actionable biological dependencies, thereby laying the groundwork for mechanism-based therapeutic strategies.