Cell-of-Origin, not Oncogenic Effect, Determines esmoplastic Immune Exclusion in KRAS-Driven Liver Cancer

This study demonstrates that in KRAS-driven liver cancer, the cell-of-origin (cholangiocyte versus hepatocyte) rather than the oncogenic driver itself dictates the formation of a desmoplastic, immune-excluded tumor microenvironment, a process mediated by cholangiocyte-specific secreted factors LAMC2 and uPA that activate cancer-associated fibroblasts.

The Gist

Imagine your liver as a bustling city with two main types of residents: Hepatocytes (the hardworking factory workers who process toxins) and Cholangiocytes (the maintenance crew who manage the plumbing and drainage pipes).

Sometimes, bad mutations happen in the DNA of these cells, turning them into cancer. The big question scientists have been asking is: Why do some liver cancers look and act completely different from others, even if they have the exact same "bad code" (mutations)?

Specifically, why is Cholangiocarcinoma (cancer of the drainage pipes) so tough to treat? It builds a thick, impenetrable wall around itself that blocks drugs and immune cells. Meanwhile, Hepatocellular Carcinoma (cancer of the factory workers) is usually more open and accessible.

This paper solves that mystery with a brilliant experiment. Here is the story in simple terms:

1. The "Same Recipe, Different Baker" Experiment

Scientists often thought the "bad code" (specifically the KRAS mutation) was the sole reason for the tough walls. To test this, they used a technique called organoids (miniature 3D liver tissues grown in a lab).

• Group A: They took the "plumbing crew" cells, gave them the bad KRAS mutation, and watched them grow.
• Group B: They took the "factory worker" cells, gave them the exact same bad KRAS mutation, and watched them grow.

The Result: Even though they had the same bad mutation, the two groups built completely different tumors.

• The Plumbing Crew built a fortress with a thick, dense wall (desmoplasia) that kept the immune system out.
• The Factory Workers built a more open city that allowed immune cells to wander in freely.

The Lesson: It's not just the bad code that matters; it's who the cell was originally. The "cell-of-origin" (the lineage) dictates the neighborhood's architecture, not just the criminal's DNA.

2. The "Immune Exclusion" Wall

The "Plumbing Crew" tumors built a massive barrier made of Cancer-Associated Fibroblasts (CAFs). Think of these CAFs as construction workers hired by the cancer to build a moat.

• In the Plumbing Crew tumors, these construction workers lined up right at the edge of the tumor, creating a dense wall of collagen and muscle.
• This wall physically blocked the body's "police force" (T-cells) from entering the tumor core.
• The immune cells were stuck outside the gate, exhausted and frustrated, while the cancer grew unchecked inside.

In the Factory Worker tumors, the construction workers didn't build a wall. The police could walk right into the tumor and do their job.

3. Finding the "Architects" (The Secret Weapons)

If the bad mutation wasn't the only cause, what was? The scientists realized the "Plumbing Crew" cells were secretly whispering instructions to the construction workers (CAFs) to build the wall.

They looked for the specific "whispers" (chemical signals) that only the Plumbing Crew cells sent out. They found two main culprits:

1. LAMC2
2. uPA

These are like specialized blueprints that only the plumbing cells possess. When these blueprints are sent out, they tell the construction workers: "Build a super-strong wall right here, and keep the police out!"

4. The "Demolition" Test

To prove this, the scientists used gene-editing tools (CRISPR) to delete the instructions for LAMC2 and uPA in the Plumbing Crew cells.

• Without these blueprints: The Plumbing Crew cells could no longer build the wall.
• The Result: When they put these "demolished" cells into mice, the tumors were much smaller, and the immune system was finally able to get inside and attack the cancer.

Why This Matters

This discovery changes how we think about cancer treatment.

• Old Thinking: "We have a KRAS mutation, so we use a KRAS drug."
• New Thinking: "We have a KRAS mutation, but where did it start? If it started in the plumbing cells, we need to also break down the wall (target LAMC2 or uPA) before the immune system can help."

The Takeaway:
Imagine trying to stop a fire. If the fire is in a wooden shed (Factory Worker cancer), a hose works fine. But if the fire is in a stone bunker (Plumbing Crew cancer), the hose can't get through the walls. You don't just need a better hose (immune therapy); you need to know what kind of building you are dealing with and bring a sledgehammer to break the walls first.

This paper tells us that the identity of the cell is the most important factor in how the cancer builds its defenses, and by targeting those specific defenses, we might finally crack the code on treating the toughest liver cancers.

Technical Summary

1. Problem Statement

Intrahepatic cholangiocarcinoma (iCCA) and hepatocellular carcinoma (HCC) are the two most common primary liver cancers. Despite sharing common risk factors and often overlapping oncogenic drivers (such as KRAS mutations and TP53 loss), they exhibit fundamentally different tumor microenvironments (TME).

• iCCA is characterized by a dense, desmoplastic stroma rich in cancer-associated fibroblasts (CAFs) that creates a physical barrier, excluding immune cells and rendering the tumor "cold" and resistant to immunotherapy.
• HCC typically displays broader immune infiltration.

The Core Question: Is this stark difference in TME architecture driven by the specific oncogenic pathways activated (e.g., KRAS signaling) or by the cell-of-origin (cholangiocyte vs. hepatocyte)? Previous models often conflated these variables (e.g., inducing KRAS in hepatocytes often leads to transdifferentiation into biliary-like cells), making it difficult to isolate the independent contributions of lineage versus genotype.

2. Methodology

The authors developed a rigorous, syngeneic, orthotopic murine model system using organoid technology to decouple cell lineage from oncogenic genotype.

• Organoid Engineering:

  • Source: Wild-type cholangiocyte and hepatocyte organoids were derived from C57BL/6 mice.
  • Genetic Modification: Both lineages were engineered to carry the same oncogenic drivers: Trp53 deletion and Kras^G12D mutation.
    • Cholangiocyte-derived (chol-PK): CRISPR-Cas9 editing.
    • Hepatocyte-derived (hep-PK): Lentiviral induction of Trp53 deletion and Kras^G12D overexpression.
  • Validation: Organoids were confirmed to retain lineage-specific markers (CK19+/HNF4α- for cholangiocytes; CK19-/HNF4α+ for hepatocytes) and were implanted orthotopically into syngeneic mice.

• Multi-Omics and Spatial Profiling:

  • Multiplexed Ion Beam Imaging (MIBI): A 30-plex antibody panel was used to profile ~390,000 cells across tumor borders and cores, mapping the spatial distribution of immune cells, CAFs, endothelial cells, and cancer cells.
  • Transcriptomics: Bulk RNA-seq was performed on all four organoid conditions (WT and PK for both lineages) to perform variance partitioning and identify lineage-specific gene expression programs.
  • Secretomics/Proteomics: Mass spectrometry (LC-MS/MS) was used to analyze the secretome (concentrated conditioned media) and proteome of the organoids to identify soluble factors driving stromal activation.
  • Functional Validation:
    • In vitro: Murine hepatic stellate cells (mHSCs) were treated with organoid conditioned media (CCM) with or without neutralizing antibodies.
    • In vivo: CRISPR-generated knockout organoids (deleting candidate genes) were implanted to assess tumor formation and TME changes.

3. Key Contributions

1. Decoupling Lineage and Genotype: The study provides definitive evidence that in a controlled setting with identical oncogenic drivers, the cell-of-origin is the dominant determinant of tumor morphology and TME architecture.
2. Identification of Lineage-Specific Mediators: The authors identified LAMC2 (Laminin subunit gamma-2) and uPA (Urokinase-type plasminogen activator) as cholangiocyte-specific secreted factors that drive CAF activation and immune exclusion.
3. Mechanistic Insight into Immune Exclusion: The study elucidates how cholangiocyte-derived tumors create a physical stromal barrier via peripherally enriched $\alpha$SMA+ myofibroblastic CAFs, whereas hepatocyte-derived tumors do not.

4. Key Results

A. Lineage Dictates Tumor Identity and TME

• Morphology: Despite sharing Trp53 loss and Kras^G12D, chol-PK tumors formed glandular, desmoplasia-rich structures resembling human iCCA, while hep-PK tumors formed trabecular, stroma-poor structures resembling HCC.
• Spatial Architecture (MIBI):
  • chol-PK (iCCA): Showed a sharp accumulation of $\alpha$SMA+ CAFs and immune cells at the tumor periphery, with a paucity of immune cells in the core. This created a "stromal barrier."
  • hep-PK (HCC): Displayed broad, sustained infiltration of immune cells and a more even distribution of CAFs throughout the tumor.
• Immune Status: chol-PK tumors exhibited higher PD-L1 expression on cancer cells and elevated PD-1 on T cells, indicating increased checkpoint engagement and immune exhaustion, consistent with an immune-excluded phenotype.

B. Transcriptional Variance Partitioning

• RNA-seq analysis revealed that lineage (cholangiocyte vs. hepatocyte) accounted for the vast majority of transcriptional variance (PC1), far outweighing the variance explained by oncogenic status (WT vs. PK) or mouse strain.
• This confirms that the oncogenic drivers modulate, but do not override, the pre-existing lineage-specific transcriptional programs.

C. Identification of Secreted Drivers (LAMC2 and uPA)

• Secretome Analysis: Integrated analysis of murine transcriptomics, secretomics, and proteomics, cross-referenced with human datasets (TCGA, CCLE), identified LAMC2, uPA (PLAU), and BSSP-4 (PRSS22) as factors specifically upregulated in cholangiocyte-derived tumors.
• Functional Validation:
  • In vitro: Conditioned media from chol-PK organoids activated HSCs (upregulating Acta2, Col1a1). Neutralizing LAMC2 or uPA significantly reduced this activation; neutralizing BSSP-4 had no effect. Recombinant uPA enhanced HSC migration.
  • In vivo: Genetic deletion of Lamc2 or Plau in chol-PK organoids resulted in a drastic reduction in tumor penetrance (17% and 33% vs. 86% in controls) and smaller tumor sizes.
  • TME Impact: Tumors that did form in the Lamc2 and Plau knockout groups showed increased T-cell infiltration, suggesting these factors are required for establishing the immune-excluded stromal barrier.

5. Significance and Implications

• Paradigm Shift: The findings challenge the assumption that oncogenic pathways (like KRAS) universally dictate TME characteristics. Instead, the cellular context (lineage) dictates how these pathways are interpreted and executed.
• Therapeutic Strategy: The study identifies LAMC2 and uPA as functional, lineage-specific modulators of the TME in iCCA. Targeting these factors could potentially disrupt the desmoplastic barrier, converting "cold" iCCA tumors into "hot" tumors amenable to immunotherapy.
• Clinical Relevance: This explains why KRAS inhibitors might have different efficacy or resistance mechanisms in HCC vs. iCCA. It suggests that cancer-agnostic therapies targeting KRAS must be combined with strategies that address the specific lineage-encoded stromal programs to be effective.
• Model Utility: The syngeneic organoid model provides a robust platform for dissecting the interplay between cell-of-origin, oncogenesis, and the tumor niche, offering a more accurate preclinical system for liver cancer research.