Transcriptomic Profiles from Normal and Tumor Tissue Samples Reveal Distinct Venule Populations and Novel Tumor Endothelial Cell Markers in Breast Cancer

This study integrates single-cell RNA-seq data to reveal that breast tumor endothelial cells exhibit an anergic phenotype with a venule predominance and distinct gene expression profiles, identifying novel markers like ADM5 that correlate with poor patient survival and resistance to immunotherapy.

The Gist

The Big Picture: The "City" of a Tumor

Imagine a breast tumor isn't just a lump of bad cells; think of it as a chaotic, growing city. For this city to grow and spread, it needs a supply network: roads, water pipes, and delivery trucks. In the body, these are the blood vessels.

In a healthy body, these "roads" (blood vessels) are well-organized, clean, and open to traffic. But in a tumor, the city planners (the cancer cells) build a mess. The roads are leaky, twisted, and clogged. Worse yet, the tumor has a secret weapon: it puts up "Do Not Enter" signs for the police (the immune system) and the delivery trucks (chemotherapy drugs).

This study is like a team of detectives using a high-tech microscope (single-cell RNA sequencing) to look at the individual "road workers" (endothelial cells) inside the tumor city versus a healthy city. They wanted to find out: How are the road workers in the tumor different from the ones in a healthy body, and can we find a way to fix the roads?

---

Key Discovery 1: The "Venule" Takeover

In a healthy breast, the blood vessels are a mix of different types: some are like small capillaries (the tiny alleys), and some are like venules (the small streets where traffic slows down).

The Finding: The researchers found that in the tumor, the mix changes completely. The tumor is almost entirely made up of venules (the small streets).

The Analogy: Imagine a healthy neighborhood has a mix of driveways, alleys, and main streets. The tumor neighborhood, however, has been paved over so that everything looks like a slow-moving, congested side street. This change isn't random; it's a specific adaptation the tumor makes to survive.

Key Discovery 2: The "Anergy" Effect (The Sleepy Guards)

One of the biggest problems with tumor blood vessels is that they are "anergic." This is a fancy word for sleepy or unresponsive.

The Finding: In a healthy body, when there is an infection or inflammation, the blood vessel walls wake up, put out "Stop" signs (adhesion molecules like SELE), and let the immune system's police cars (white blood cells) pull over and enter the tissue to fight the bad guys.

In the tumor, the blood vessels are asleep. They have taken down the "Stop" signs. They are ignoring the alarm bells. Because of this, the immune system's police cars drive right past the tumor without ever getting out of their cars to fight the cancer.

The Analogy: Think of the tumor blood vessels as a sleeping security guard at a bank. Even though the alarm is ringing (inflammation), the guard is asleep and won't let the police (immune cells) inside. The cancer thrives because no one is coming to stop it.

Key Discovery 3: The "Bad Signaling" (NF-kB)

Why are the guards asleep? The researchers found that a specific signaling pathway inside the cells (called NF-kB) is broken.

The Analogy: In a healthy cell, NF-kB is like the fire alarm system. When there's trouble, it rings loud and clear, waking everyone up. In the tumor cells, someone has cut the wires to the alarm. The alarm (inflammatory signals) isn't ringing, so the cells stay asleep.

The New "ID Badges" (Novel Markers)

The researchers wanted to find a way to target only the tumor's blood vessels without hurting the healthy ones. To do this, they looked for unique "ID badges" (proteins) that the tumor cells wear but normal cells don't.

They found four potential badges:

1. CLEC14a and EMCN: These were already known to be on tumor cells, but the study confirmed they are very common in breast cancer.
2. IGFBP4: This one was a bit messy; it appeared on other cells too, so it's not a perfect target.
3. ADM5: This is the big new discovery.

The ADM5 Analogy: Imagine ADM5 is a glowing red hat that only the tumor's road workers wear. Normal road workers don't wear it.

• It is found mostly on the "venule" streets (the ones the tumor loves).
• It is very rare in healthy tissue.
• The Bad News: Patients with high levels of this "red hat" (ADM5) tend to have shorter survival times, especially in aggressive types of breast cancer.
• The Good News: Because it's unique to the tumor, it's a perfect target for a "smart bomb" therapy. We could design a drug that specifically targets the red hat, waking up the sleepy guards or destroying the bad road workers, without hurting the healthy city.

The Immunotherapy Connection

The study also looked at how this affects modern cancer treatments called immunotherapy (drugs like PD-1 or CTLA-4 inhibitors that try to wake up the immune system).

The Finding: Patients with high levels of the "red hat" (ADM5) did worse when treated with PD-1 or CTLA-4 drugs. However, they did okay with a different drug (PDL-1).

The Analogy: It's like the tumor's "red hat" has a special shield that blocks the PD-1 police officers from entering. But if you use a different type of police officer (PDL-1), they can still get in. This tells doctors that if a patient has high ADM5, they might need a different treatment plan to be effective.

The Conclusion: What's Next?

This paper is a map. It tells us:

1. Tumor blood vessels are different from healthy ones (they are mostly "venules").
2. They are "sleepy" (anergic) because their alarm systems are broken.
3. They wear a unique "red hat" called ADM5.

The Takeaway: If we can develop drugs that target ADM5, we might be able to:

• Wake up the sleepy blood vessels.
• Let the immune system's police cars finally enter the tumor.
• Stop the tumor from growing and spreading.

It's a hopeful step toward turning the tumor's own defenses against it, using the unique "ID badges" the cancer cells accidentally left behind.

Technical Summary

1. Problem Statement

The breast tumor microenvironment (TME) is a complex system where aberrant tumor vasculature contributes significantly to therapeutic resistance, poor drug delivery, and immune evasion. A critical phenomenon in this context is endothelial anergy, a state where tumor endothelial cells (TECs) become unresponsive to inflammatory signals, downregulating adhesion molecules (e.g., SELE, ICAM1) and failing to recruit immune cells.

Despite the known heterogeneity of TECs, there is a lack of comprehensive understanding regarding:

• The specific phenotypic shifts between normal endothelial cells (NEC) and TECs in breast cancer.
• The distribution of vascular subtypes (arterioles, capillaries, venules) in tumors versus normal tissue.
• The identification of TEC-specific markers that are not present in normal tissue, which could serve as targets for reactivating the vasculature or delivering therapies without systemic toxicity.
• The role of specific signaling pathways (e.g., NF-κB) in driving the anergic state.

2. Methodology

The study employed a rigorous bioinformatics approach using publicly available single-cell RNA sequencing (scRNA-seq) datasets.

• Data Acquisition & Selection:
  • Reviewed 11 scRNA-seq datasets (2020–2023) from the NCBI GEO.
  • Selected two high-quality, treatment-naïve datasets: GSE176078 (tumor) and GSE161529 (normal and tumor, though tumor samples were excluded due to low EC counts).
  • Criteria: >10 patients, diverse subtypes, whole-tissue analysis (no pre-enrichment), and primary tumors only.
• Data Processing & Integration:
  • Used Seurat (R package) for data integration and analysis.
  • Quality Control: Filtered cells based on mitochondrial gene expression (<10%), gene counts (200–20,000), and transcript counts.
  • Cell Type Isolation: Applied a strict "negative selection" strategy to remove non-endothelial cells (PTPRC, ACTA2, COL1A1, PDGFRB) and "positive selection" using canonical markers PECAM1 and CLDN5 to isolate endothelial cells (ECs).
  • Clustering: Performed dimensionality reduction (PCA, UMAP) and K-nearest neighbor (KNN) clustering, resulting in 8 distinct EC clusters.
• Subtype Annotation:
  • Annotated clusters based on canonical markers: Arterioles (EFNB2, HEY1), Capillaries (RGCC, KDR), Venules (ACKR1, SELP), and Lymphatics (PROX1, LYVE1).
• Differential Expression & Enrichment:
  • Used a pseudobulk approach (AggregateExpression) to compare NEC vs. TEC, reducing false positives.
  • Identified Differentially Expressed Genes (DEGs) using the MAST package with Bonferroni correction.
  • Performed Gene Ontology (GO) and Gene Set Enrichment Analysis (GSEA) using MSigDB hallmark sets.
• Validation & Clinical Correlation:
  • Specificity: Validated candidate markers in the Curated Cancer Cell Atlas (3CA) across multiple cancer types.
  • Survival Analysis: Correlated gene expression with Overall Survival (OS) and Progression-Free Survival (PFS) using Kaplan-Meier Plotter (GSE96058) and cBioPortal (TCGA/METABRIC).
  • Immunotherapy Response: Analyzed survival outcomes in patients treated with anti-PD1, anti-PDL1, and anti-CTLA4 therapies.

3. Key Contributions

• Discovery of Venule Dominance in Tumors: Identified a significant shift in vascular composition, where breast tumors are dominated by venule subtypes (specifically activated post-capillary venules), whereas normal tissue maintains a balanced distribution of capillaries and venules.
• Characterization of Endothelial Anergy: Provided transcriptomic evidence of NF-κB pathway dysregulation in TECs, characterized by the suppression of pro-inflammatory cytokines (IL6, CCL2, CXCL8) and adhesion molecules (SELE).
• Identification of Novel TEC Markers: Discovered and validated CLEC14A, EMCN, and a novel marker, ADM5, as highly specific to tumor endothelial cells, particularly in venule clusters.
• Clinical Relevance of ADM5: Established ADM5 (Adrenomedullin 5) as a prognostic biomarker associated with poor survival in basal-like breast cancer and reduced efficacy of PD1/CTLA4 immunotherapies.

4. Key Results

A. Vascular Subtype Distribution

• Normal Tissue (NEC): Balanced proportions of capillary and venule subtypes. The majority of normal venules belonged to the venule_3 cluster (quiescent phenotype).
• Tumor Tissue (TEC): Predominantly composed of venule subtypes (>92% in venule_1 and venule_2 clusters). These clusters exhibited an activated post-capillary venule phenotype.
• Statistical Significance: The distribution of EC subtypes differed significantly between normal and tumor tissues (Chi-squared test, p = 0.0246).

B. Transcriptomic Shifts and Endothelial Anergy

• DEG Analysis: 1,188 DEGs were identified. TECs showed upregulation of angiogenic and ECM remodeling genes (e.g., ARHGDIB, RAMP2, JAM2).
• Anergy Signature: TECs showed significant downregulation of inflammatory mediators compared to NECs, including SELE, ICAM1, IL6, CCL2, and CXCL8.
• NF-κB Signaling: GSEA and specific gene analysis revealed that NF-κB target genes were repressed in TECs, suggesting aberrant NF-κB signaling drives the anergic state.

C. Novel Tumor Endothelial Markers (TEMs)

Four genes were identified as upregulated in TECs: CLEC14A, EMCN, IGFBP4, and ADM5.

• CLEC14A & EMCN: Highly specific to endothelial cells across multiple cancer types (validated in 3CA).
• IGFBP4: Expressed in endothelial cells but also in fibroblasts and malignant cells; less specific.
• ADM5: A novel marker highly specific to TECs (absent in NEC). It was selectively enriched in the venule_1 and venule_2 clusters.

D. Clinical and Therapeutic Implications

• Survival Correlation:
  • High ADM5 expression correlated with significantly reduced Overall Survival (OS) in basal-like breast cancer (HR = 2.27, p = 0.0001) and other cancers (gastric, renal, ovarian).
  • High expression of RAMP2 (a receptor component for ADM5 signaling) also correlated with poor OS in basal breast cancer.
• Immunotherapy Resistance:
  • High ADM5 expression was associated with reduced PFS and OS in patients treated with anti-PD1 and anti-CTLA4 therapies.
  • No such negative correlation was observed with anti-PDL1 therapy, suggesting ADM5 specifically impacts T-cell mediated checkpoint inhibition, likely via endothelial immune evasion mechanisms.

5. Significance

This study provides a high-resolution map of the breast tumor vasculature, moving beyond bulk analysis to reveal specific venule-driven heterogeneity in tumors. The identification of ADM5 as a novel, TEC-specific marker with strong prognostic value offers a new avenue for:

1. Therapeutic Targeting: Developing agents to target ADM5 or its signaling pathway (via RAMP2/CALCRL) to disrupt tumor vascular support.
2. Reactivating Immunity: Understanding how ADM5 contributes to endothelial anergy could lead to strategies that "wake up" the tumor vasculature, enhancing immune cell infiltration and improving the efficacy of immunotherapies (specifically PD1/CTLA4).
3. Biomarker Development: Using ADM5 expression to stratify patients who may be resistant to current immunotherapies.

The findings underscore the importance of targeting the tumor endothelium not just for anti-angiogenesis, but as a strategy to overcome immune evasion and therapeutic resistance in breast cancer.