RUNX1-deficiency drives immune-active ER+ mammary tumorigenesis through activation of interferon signaling

This study demonstrates that the loss of RUNX1 drives immune-active ER+ breast tumorigenesis by activating interferon signaling and recruiting immune cells, a mechanism confirmed by both genetically engineered mouse models and human clinical data linking low RUNX1 expression to elevated immune signatures and poorer patient survival.

The Gist

The Big Picture: A Broken "Brake" and a Loud Alarm

Imagine your body is a bustling city, and your breast tissue is a specific neighborhood made up of specialized workers called Luminal Cells. These workers have a very important job: they follow a strict rulebook to know when to grow and when to stop.

In this study, scientists discovered what happens when one of the most important rulebook editors, a protein called RUNX1, goes missing.

Usually, RUNX1 acts like a strict librarian. It sits in the cell's control room, making sure the "noise" (immune signals) stays quiet and the workers don't get confused. But when RUNX1 is lost (due to a mutation), the library goes chaotic.

The Experiment: Building a Better Mouse Model

The scientists wanted to see what happens when these "librarians" go missing in breast cells. In the past, trying to study this in mice was like trying to fix a leaky roof while the whole house was on fire. When they tried to remove RUNX1 along with another famous tumor suppressor (p53), the mice got sick from blood cancers before they could even develop breast tumors.

So, the team built a specialized "smart home" mouse model.

• The Trick: They used a special virus (like a targeted delivery drone) that only unlocks the door to the breast cells.
• The Result: They could now remove RUNX1 only in the breast cells, leaving the rest of the body (and the blood system) perfectly healthy. This allowed them to watch the breast tumors grow from the very beginning.

The Discovery: Two Different Types of Chaos

The scientists tested three scenarios:

1. Missing RUNX1 alone: Nothing happened. The cells were confused but didn't turn into cancer.
2. Missing RUNX1 + Missing RB1 (another guard): Still nothing. The cells didn't grow out of control.
3. Missing RUNX1 + Missing p53 (the main "stop" button): BINGO. This combination caused aggressive tumors to grow in 100% of the mice.

The Twist: These tumors looked very different from typical breast cancers.

• Typical p53-loss tumors: These are like a factory running wild, churning out cells rapidly (high proliferation). They are "cold" and quiet.
• RUNX1-loss tumors: These were "hot." They were filled with immune cells (T-cells and macrophages), like a neighborhood under constant siege by police and firefighters. The tumors also had strange features, like squamous patches (skin-like cells), which suggested a specific type of cellular confusion.

The Mechanism: The "Silence Button" Breaks

Why were these tumors so "hot" with immune cells?

The scientists found that RUNX1 usually acts as a silence button for a specific alarm system called the Interferon (IFN) pathway.

• Normal Cell: RUNX1 is present. It keeps the "Interferon" alarm muted. The cell is quiet.
• RUNX1-Loss Cell: The silence button is broken. The alarm goes off!
  • The cell starts shouting, "Something is wrong!" by releasing chemical signals.
  • This signal is actually a gene called STAT1. Without RUNX1 to hold it back, STAT1 goes into overdrive.
  • This shouting attracts the body's immune system (the T-cells and macrophages) to the scene.

The Analogy: Imagine a house where the smoke detector is broken. Instead of just beeping once, it screams continuously. The fire department (immune system) rushes over. At first, this seems good (the body is trying to fight the cancer), but eventually, the constant noise and the presence of the fire department create a chaotic environment that actually helps the cancer survive and grow stronger.

The Human Connection: Why This Matters for Patients

The researchers checked human breast cancer data to see if this happened in real people.

• They found that women with ER+ breast cancer (the most common type, which is usually treated with hormone therapy) who had low levels of RUNX1 had a much worse outlook.
• These patients' tumors were "immune-hot," just like the mice. They had high levels of STAT1 and were flooded with immune cells.
• Paradoxically, having a "hot" immune environment usually sounds good for cancer treatment, but in this specific case, it seemed to correlate with the cancer being more aggressive and harder to treat.

The Takeaway

This paper tells us that RUNX1 is a tumor suppressor specifically for the most common type of breast cancer (ER+).

1. When RUNX1 is lost, it doesn't just let cells grow; it turns on a loud, chronic inflammatory alarm.
2. This alarm attracts immune cells, creating a "hot" tumor environment.
3. This specific type of cancer is distinct from other ER+ cancers and might require different treatments.

The Future Hope:
Because we now know these tumors are "immune-hot," doctors might be able to treat them differently. Instead of just using hormone therapy, they might combine it with immunotherapy (drugs that help the immune system fight cancer) or anti-inflammatory drugs to calm down the chaotic alarm system.

In short: Losing RUNX1 turns a quiet neighborhood into a noisy, chaotic war zone, and understanding the noise helps us figure out how to stop the war.

Technical Summary

1. Problem Statement

• Context: Recurrent loss-of-function mutations in the transcription factor RUNX1 are frequently observed in human estrogen receptor-positive (ER+) breast cancers.
• Knowledge Gap: Despite the high frequency of these mutations, the causal role of RUNX1 loss in breast tumorigenesis and the specific molecular mechanisms driving tumor formation remain unclear. Previous attempts to model this using MMTV-Cre (which has leaky expression in hematopoietic cells) resulted in early lethality due to hematopoietic malignancies, preventing the study of mammary-specific tumorigenesis.
• Objective: To establish a causal link between RUNX1 deficiency and breast cancer initiation, determine its cooperative partners (e.g., p53, RB1), and characterize the resulting tumor phenotype and microenvironment.

2. Methodology

The authors employed a combination of genetically engineered mouse models (GEMMs), intraductal viral delivery, and multi-omics analyses:

• Mouse Models:
  • Generated conditional knockout mice: Runx1^L/L, Trp53^L/L, Rb1^L/L, and Rosa26-LSL-YFP (R26Y).
  • Created specific cohorts: RxPY (Runx1/p53 double KO), RRY (Runx1/Rb1 double KO), RbPY (Rb1/p53 double KO), PRRY (Runx1/Rb1/p53 triple KO), and single KOs.
• Tumor Induction & Lineage Tracing:
  • Utilized intraductal injection of an adenovirus expressing Cre under the Keratin 8 (K8) promoter (Ad-K8-Cre). This restricts gene deletion strictly to luminal mammary epithelial cells (MECs), avoiding hematopoietic toxicity.
  • Used YFP (R26Y) for lineage tracing to track the fate of deleted MECs via Fluorescence-Activated Cell Sorting (FACS).
• Characterization Techniques:
  • Histology: H&E staining and immunofluorescence (IF) for markers (ERα, K8, β-catenin, LEF1, Ki67, pSTAT1).
  • Flow Cytometry (FACS): Multi-color analysis to quantify immune cell infiltration (CD45+, CD3+, CD4/CD8, F4/80+ macrophages) in both premalignant glands and tumors.
  • Transcriptomics: RNA-seq and microarray analysis of sorted YFP+ MECs and whole tumors.
  • Bioinformatics: Gene Set Enrichment Analysis (GSEA) using MSigDB Hallmark and immune signatures; comparison with human datasets (METABRIC and TCGA).
  • Mechanistic Validation: ChIP-seq analysis of RUNX1 binding sites and correlation studies in human cell lines (MCF7, MCF10A) with RUNX1 knockdown/overexpression.

3. Key Contributions & Results

A. Genetic Cooperation and Tumorigenesis

• Cooperation with p53: Loss of Runx1 alone or combined with Rb1 was insufficient to drive tumors. However, combined loss of RUNX1 and p53 (RxPY model) induced mammary tumors with 100% penetrance.
• Specificity: Combined loss of Runx1 and Rb1 (RRY) did not cause tumors, even with p53 loss (PRRY triple KO), though PRRY tumors resembled RxPY tumors. This indicates Runx1 specifically cooperates with p53, not Rb1, in this context.
• Latency: RxPY tumors had longer latency than RbPY (Rb1/p53 KO) tumors, suggesting a different biological mechanism (e.g., immune surveillance) slowing progression.

B. Tumor Phenotype and Identity

• ER+ Status: RxPY and PRRY tumors were predominantly ERα-positive, consistent with the human clinical observation that RUNX1 mutations occur in ER+ breast cancer.
• Histology: These tumors exhibited squamous metaplasia and adenocarcinoma features, distinct from the poorly differentiated Claudin-low tumors seen in p53-only models.
• Wnt/β-catenin Activation: The squamous phenotype correlated with nuclear accumulation of β-catenin and LEF1, suggesting Runx1 loss aberrantly activates Wnt/β-catenin signaling.

C. Immune Microenvironment ("Immune Hot")

• Infiltration: RxPY and PRRY tumors displayed extensive infiltration of CD45+ leukocytes, specifically CD3+ T cells and F4/80+ macrophages, compared to RbPY controls.
• Premalignant Stage: Increased T cell infiltration (specifically CD8+ initially, shifting to CD4+ dominance later) was detected as early as 2 weeks post-induction in premalignant glands, indicating the immune response is triggered by the epithelial defect itself, not just secondary to tumor bulk.
• Human Correlation: Human ER+ breast cancers with low RUNX1 expression (METABRIC/TCGA data) showed significantly higher immune signatures (T cells/macrophages) and poorer overall survival.

D. Molecular Mechanism: Interferon (IFN) Signaling

• IFN Activation: RNA-seq of Runx1-deficient premalignant MECs revealed strong enrichment of Interferon (IFN) response gene sets. This was not observed in p53-only or Brca1/p53-deficient cells at the same early stage.
• STAT1 Derepression:
  • ChIP-seq identified RUNX1 binding sites in the STAT1 locus (promoter and intronic regions) within repressive chromatin valleys.
  • Loss of RUNX1 led to upregulation of STAT1 mRNA and protein (pSTAT1) in both mouse models and human cell lines (MCF7).
  • A negative correlation between RUNX1 and STAT1 expression was confirmed in human ER+ tumors.
• Mechanism: RUNX1 normally represses STAT1. Its loss leads to STAT1 overexpression, enhancing IFN signaling and recruiting immune cells.

4. Significance

• Novel Driver Mechanism: Establishes RUNX1 loss as a bona fide driver of a distinct subtype of ER+ breast cancer, distinct from p53-only or Rb1-driven tumors.
• Immune-Active ER+ Subtype: Challenges the paradigm that ER+ breast cancers are generally "immune cold." It identifies a specific subset driven by RUNX1 loss that is "immune hot" due to intrinsic epithelial IFN signaling.
• Clinical Implications:
  • Prognosis: RUNX1-low/STAT1-high ER+ tumors are associated with worse patient survival.
  • Therapeutic Strategy: These tumors may be uniquely susceptible to immunotherapy (e.g., immune checkpoint blockade) or combination therapies targeting the IFN pathway, offering new avenues for treating resistant ER+ breast cancers.
• Species Conservation: The link between RUNX1 loss, STAT1 upregulation, and immune activation is conserved between murine models and human breast cancer datasets.

Conclusion

The study demonstrates that RUNX1 deficiency in luminal mammary epithelial cells, particularly when combined with p53 loss, drives tumorigenesis by derepressing STAT1, activating interferon signaling, and creating an immune-active microenvironment. This defines a distinct, aggressive subtype of ER+ breast cancer with significant implications for prognosis and immunotherapy development.