Estrogen receptor-positive cell line xenograft models recapitulate metastatic dissemination and endocrine response of invasive lobular breast carcinoma

This study demonstrates that reporter-labeled estrogen receptor-positive invasive lobular breast carcinoma xenografts faithfully recapitulate the unique single-file histology, specific metastatic patterns (including brain and bone), and robust endocrine therapy response of human disease, establishing them as valuable pre-clinical platforms for therapeutic development.

The Gist

Imagine the human body as a bustling city. In this city, breast cancer is like a group of rebellious cells that decide to stop following the rules and start taking over. Most of the time, these rebels are "Invasive Ductal Carcinoma" (IDC), which is the most common type. But there's a specific, tricky subgroup called Invasive Lobular Carcinoma (ILC).

Think of ILC as a "ghost" or a "sneaky spy" compared to the other rebels.

• The Look: While normal cancer cells stick together like a tight-knit gang, ILC cells lose their "glue" (a protein called E-cadherin). Instead of forming a solid lump, they spread out in single-file lines, like a long queue of people walking through a hallway. This makes them hard to spot on standard scans.
• The Destination: Regular cancer usually goes to the lungs or liver. But these "ghosts" have a weird habit of traveling to unusual places like the ovaries, the lining of the brain, the eyes, and the gut.
• The Fuel: Almost all of them run on Estrogen, a natural hormone in the body. Doctors usually treat them by cutting off this fuel supply (endocrine therapy).

The Problem: The "Fake" Models

For years, scientists tried to study ILC in mice to find better cures. They built "genetically engineered" mice (like building a fake city from scratch to test traffic). But these fake cities didn't quite work. They missed the "ghostly" single-file spread, they didn't go to the weird places (like the ovaries), and they didn't respond to the hormone-blocking drugs the way real human patients do. It was like testing a car on a track that didn't have any potholes or traffic lights.

The Solution: The "Real-World" Simulation

This paper introduces a new, much better way to study ILC. Instead of building a fake city, the scientists took real human ILC cells (the actual rebels) and injected them into special "immunocompromised" mice (mice with a disabled security system so they don't reject the human cells).

Here is what they discovered, using some creative analogies:

1. The Perfect Mimicry

When they injected these human cells into the mice's breast tissue (the "mammary fat pad"), the tumors grew slowly and formed those classic single-file lines.

• Analogy: It was like dropping a real seed into the soil, and it grew into the exact same weird plant as the one in the wild, rather than a plastic fake.
• The Result: These mouse tumors looked, acted, and spread exactly like the human disease. They even sent "spies" to the ovaries, brain, and bones, just like real patients.

2. The "Ghost" in the Brain

One of the scariest things about ILC is that it can hide in the brain, specifically in the thin lining around it (the leptomeninges).

• The Discovery: The scientists watched these mouse models and saw the cancer cells creeping along the brain's lining, just like in humans. They even saw the cancer eating away at the bone (like termites eating wood).
• Why it matters: This proves the model is accurate enough to study how the cancer hides in the brain, which is a major cause of death.

3. The "Fuel Cut-Off" Test

Since ILC runs on estrogen, doctors try to stop the fuel.

• The Experiment: The scientists gave the mice a drug called Fulvestrant (which acts like a "fuel cap" that blocks estrogen).
• The Outcome: The tumors in the mice shrunk and stopped growing. Even the cancer that had already spread to the brain and bones responded to the treatment.
• The Metaphor: It's like finding a car that runs on a specific type of gas, then cutting off the gas line, and watching the car sputter and stop. This confirms that the mouse model reacts to medicine just like a human patient would.

4. The "Smoke Detector" (Tracking the Disease)

The scientists also tested a new way to track the cancer. They looked for "cell-free DNA" in the mice's blood.

• Analogy: Imagine the cancer cells are burning a fire. Even if you can't see the fire, you can smell the smoke. The scientists found "smoke" (DNA fragments from the cancer) in the blood, proving they could detect the cancer spreading even before it became a big lump.

The Big Picture

This paper is a breakthrough because it finally gives scientists a reliable "test drive" for ILC treatments.

Before, testing new drugs was like trying to fix a car using a toy model that didn't have the same engine. Now, they have a "real" model that behaves exactly like the human disease.

• For the future: This means scientists can test new drugs on these mice with confidence. If a drug works on the mouse "ghosts," it has a much better chance of working on human patients. It also helps them understand why the cancer goes to the brain or ovaries, which could lead to new ways to stop it there.

In short: The scientists built a perfect "virtual reality" simulation of Invasive Lobular Carcinoma using real human cells in mice. This simulation is so accurate that it mimics the disease's sneaky spread, its weird destinations, and its response to treatment, giving hope for better cures for patients.

Technical Summary

1. Problem Statement

Invasive Lobular Carcinoma (ILC) accounts for 10–15% of all breast cancer cases and is characterized by the loss of E-cadherin (CDH1), leading to a unique single-file histological growth pattern. Unlike Invasive Ductal Carcinoma (IDC), ILC exhibits distinct metastatic patterns, frequently disseminating to unusual sites such as the ovaries, peritoneum, gastrointestinal tract, and leptomeninges (brain coverings), in addition to bone and brain.

Current preclinical models, including Genetically Engineered Mouse Models (GEMMs) and existing Patient-Derived Xenografts (PDXs), have limitations:

• They often fail to fully recapitulate the unique histology (single-file growth) and specific metastatic tropism (e.g., ovarian and leptomeningeal spread) of human ILC.
• Many models lack robust estrogen receptor (ER) signaling or do not faithfully mimic the clinical response to endocrine therapy, which is the standard of care for ER+ ILC.
• There is a need for comprehensive, reproducible xenograft models that can serve as platforms for validating genetic drivers and testing novel therapeutics for ILC.

2. Methodology

The authors developed and characterized a suite of orthotopic and experimental xenograft models using four ER-positive human ILC cell lines: MDA-MB-134-VI, SUM44PE, MDA-MB-330, and BCK4.

• Cell Line Engineering: Cell lines were engineered to express Luciferase and tRFP (turbo Red Fluorescent Protein) via lentiviral infection to enable non-invasive bioluminescent imaging (BLI) and fluorescence tracking.
• Orthotopic Xenografts (Spontaneous Metastasis):
  • 5 million cells were injected into the mammary fat pads (MFP) of immunocompromised NSG mice (NOD-scid IL2Rgamma null).
  • Mice were supplemented with estradiol pellets to support ER+ tumor growth.
  • Primary tumors were monitored via caliper measurements and BLI. In some cohorts, primary tumors were surgically resected to test for early dissemination.
• Experimental Metastasis Models:
  • Tail Vein Injection: Cells injected into the tail vein of NSG and athymic nude mice to assess lung colonization and systemic spread.
  • Intracardiac Injection: Cells injected into the left ventricle of nude mice to model arterial dissemination and bone metastasis.
• Metastasis Characterization:
  • Imaging: Longitudinal in vivo BLI and ex vivo fluorescence imaging of harvested organs.
  • Histology: H&E staining, Immunohistochemistry (IHC) for ER, PR, E-cadherin (CDH1), p120-catenin, PDGFRβ, and Ly6G. Masson's Trichrome staining for collagen deposition.
  • Leptomeningeal Analysis: Specific protocols to identify and stage brain metastases (early vs. late) using CK19 staining to confirm human origin.
  • Liquid Biopsy: Circulating cell-free DNA (cfDNA) was extracted from mouse plasma and analyzed via Digital Droplet PCR (ddPCR) to detect specific mutations (KRAS G12R and MAP2K4 S184L) present in the MDA-MB-134 line.
  • Transcriptomics: Bulk RNA-Sequencing (RNA-seq) was performed on primary tumors and spontaneous brain metastases. Mouse reads were computationally removed using XenoFilteR to isolate human tumor transcriptomes. Differential gene expression (DGE) and Gene Set Enrichment Analysis (GSEA) were conducted.
• Therapeutic Intervention: Mice with established tumors were treated with Fulvestrant (a Selective Estrogen Receptor Degrader, SERD) or vehicle to assess endocrine response in both primary and metastatic settings.

3. Key Contributions

• Development of Robust ILC Xenografts: The study successfully established orthotopic models using four distinct ILC cell lines that faithfully mimic the slow growth, single-file histology, and E-cadherin loss of human ILC.
• Recapitulation of Unique Metastatic Tropism: The models demonstrated spontaneous metastasis to clinically relevant but rare ILC sites, including ovaries, adrenal glands, and leptomeninges, which are difficult to model in other systems.
• Leptomeningeal Metastasis (LM) Model: The authors provided a detailed characterization of LM formation, showing that ILC cells can invade the leptomeninges and form macrometastases, a critical unmet need in ILC research.
• Endocrine Responsiveness: The models were validated as highly responsive to Fulvestrant, demonstrating significant tumor regression and survival benefits, confirming their utility for testing endocrine therapies.
• cfDNA Tracking: The study proved the feasibility of monitoring metastatic burden and specific driver mutations (MAP2K4, KRAS) in these models via circulating tumor DNA (ctDNA), bridging the gap to clinical liquid biopsy applications.

4. Key Results

• Tumor Growth and Histology: Orthotopic tumors exhibited slow growth kinetics (with delayed onset changepoints for BCK4 and MDA-MB-330) and displayed the classic single-file growth pattern with stromal collagen deposition. All lines retained high ER expression and lost E-cadherin with cytoplasmic p120 relocalization.
• Spontaneous Metastasis:
  • Tumors spontaneously metastasized to bone, brain, ovaries, and adrenal glands.
  • Leptomeningeal Metastasis: MDA-MB-134 and SUM44PE cells formed LM lesions. Early stages showed scattered cells in the leptomeninges, while late stages featured macrometastases near the pons and ventral midbrain.
  • Bone Metastasis: Micro-CT confirmed lytic bone lesions with erosion, mimicking clinical ILC bone metastases.
  • Primary Tumor Independence: Surgical removal of primary tumors did not stop metastatic progression, suggesting early dissemination occurs before the primary tumor is detectable.
• Experimental Metastasis:
  • Tail Vein: MDA-MB-134 cells generated widespread metastases in NSG mice (adrenal, ovary, brain, bone) but notably avoided lung colonization, contrasting with typical lung tropism in other models. This suggests the primary tumor may "educate" the lung niche, a phenomenon absent in direct tail vein injection.
  • Intracardiac: MDA-MB-134 cells successfully colonized bone, eye, and brain, whereas SUM44PE failed to establish detectable metastases via this route.
• Transcriptomic Profiling:
  • RNA-seq of brain metastases vs. primary tumors revealed distinct gene expression profiles.
  • GSEA showed enrichment of Estrogen Signaling (both early and late response) in brain metastases.
  • Upregulation of RET (an estrogen-regulated tyrosine kinase) and downregulation of BLNK were observed, suggesting elevated ER activity in the brain microenvironment.
• Therapeutic Response:
  • Fulvestrant treatment significantly inhibited primary tumor growth and reduced metastatic burden (brain, ovary, lung) in the orthotopic setting.
  • In the intracardiac (established metastasis) setting, Fulvestrant reduced metastatic signal but did not significantly extend survival, highlighting the aggressive nature of established metastatic disease.
• cfDNA Detection: ddPCR successfully detected tumor-specific mutations (MAP2K4 and KRAS) in the plasma of mice bearing both orthotopic and tail vein xenografts, correlating with metastatic burden.

5. Significance

This study provides a comprehensive set of preclinical tools that address critical gaps in ILC research:

1. Faithful Modeling: The models overcome the limitations of GEMMs by recapitulating the specific metastatic tropism (ovary, leptomeninges) and endocrine sensitivity of human ILC.
2. Drug Development: The demonstrated responsiveness to Fulvestrant validates these models for screening novel endocrine therapies and combination strategies.
3. Mechanistic Insights: The identification of ER signaling enrichment in brain metastases and the potential role of RET offers new therapeutic targets for ILC brain metastases.
4. Translational Relevance: The successful use of cfDNA tracking in these models supports the development of liquid biopsy biomarkers for monitoring ILC progression and treatment response in clinical trials.

Collectively, these xenograft models serve as a vital platform for the global research community to better understand ILC biology and accelerate the development of effective treatments for this distinct and challenging subtype of breast cancer.