Inflammatory memory disables a DDR1 degradation checkpoint to enable pancreatic cancer growth

This study reveals that inflammatory signaling establishes a "memory" by silencing the FBXW2-mediated degradation of the collagen receptor DDR1, thereby allowing pancreatic cancer cells to sustain growth-promoting signaling even in metabolically restrictive stromal environments.

The Gist

The Big Picture: A Cancer Survival Story

Imagine Pancreatic Cancer (PDAC) as a tough, stubborn intruder trying to build a fortress in a very crowded, hostile neighborhood. This neighborhood is the tumor microenvironment, which is packed with a thick, rigid mesh of collagen (a type of structural protein, like steel beams in a building).

Usually, this thick mesh is supposed to crush the cancer. But this cancer has learned a secret trick: it can turn the "steel beams" of the neighborhood into "fuel" to keep growing. This paper discovers how it does that, and more importantly, how it remembers this trick even when the neighborhood changes.

---

1. The Two Types of "Beams" (Collagen)

Think of the collagen in the tumor as two different types of wood:

• Intact Collagen (The Solid Log): This is the original, uncut, rigid beam. It's hard and unyielding. When the cancer cells touch this, they can't get a good grip. They feel weak, their energy (ATP) drops, and they start to shut down.
• Cleaved Collagen (The Shavings): Enzymes in the tumor chop the solid logs into smaller shavings (fragments). These shavings are like high-octane fuel. When the cancer cells grab these shavings, they get a massive energy boost and start growing wildly.

The Problem: If the tumor can't chop the logs into shavings, it should die. But sometimes, it doesn't. Why?

2. The "Off Switch" (DDR1 and FBXW2)

The cancer cell has a special receptor on its surface called DDR1. Think of DDR1 as a gas pedal.

• When the cell grabs the "shavings" (cleaved collagen), the gas pedal is pressed down, and the car (tumor) speeds up.
• When the cell touches the "solid log" (intact collagen), the gas pedal doesn't work well. The car stalls.

To prevent the car from stalling completely, the cell has a safety mechanism. If the gas pedal isn't working (because there are no shavings), the cell has a mechanic named FBXW2.

• FBXW2's Job: It sees that the gas pedal (DDR1) is useless because there's no fuel. So, FBXW2 grabs the gas pedal and throws it in the trash (degrades it). This saves energy. The car stops running because it has no gas and no pedal.

3. The "Energy Alarm" (AMPK)

How does FBXW2 know to throw away the gas pedal? It listens to the car's fuel gauge.

• When the cell is touching the "solid log," its energy (ATP) drops.
• This triggers an Energy Alarm called AMPK.
• The alarm tells FBXW2: "Hey, we have no fuel! The gas pedal is broken! Throw it away!"
• FBXW2 phosphorylates (tags) the DDR1 gas pedal and sends it to the trash. The tumor stops growing.

4. The "Bad Memory" (Inflammatory Memory)

Here is the twist. The paper discovers that inflammation (caused by cytokines like IL-8 or IL-17) acts like a hacker.

• The Hack: The inflammation tells the cell's "security system" to lock the door on the FBXW2 mechanic. It does this by putting a heavy padlock (methylation) on the FBXW2 gene.
• The Result: The FBXW2 mechanic is silenced. It can't get into the room to throw away the gas pedal.
• The Memory: Even if the inflammation goes away, the padlock stays on. The cell has an "Inflammatory Memory." It remembers the inflammation and keeps the FBXW2 door locked forever.

5. The Ultimate Cheat Code

Because the FBXW2 mechanic is locked out, the DDR1 gas pedal stays on the dashboard, even if the cell is touching the "solid log" (intact collagen) and has no fuel.

• The cell tricks itself into thinking it has fuel.
• It keeps the gas pedal pressed down.
• It keeps the engine running (growing and spreading) even in the harshest, most restrictive environments where it should have died.

Summary: The Analogy in a Nutshell

1. Normal Situation: The tumor is in a tough neighborhood (solid collagen). It runs out of gas. A mechanic (FBXW2) sees this, removes the gas pedal (DDR1), and the tumor stops.
2. The Hack: The tumor gets exposed to inflammation (like a bad neighborhood riot). This riot tells the tumor to padlock the mechanic's door.
3. The Cheat: Even after the riot is over, the door is still locked. The mechanic can't get in. The gas pedal stays attached. The tumor keeps driving fast, even though it's running on empty and surrounded by solid walls.

Why This Matters

This discovery explains why pancreatic cancer is so hard to kill. It's not just about the physical environment; it's about the cancer's memory of past inflammation.

• The Good News: We now know exactly how to break this cycle. If we can unlock the door (remove the methylation padlock) or force the mechanic back to work, we can force the cancer to stop growing, even if it's surrounded by tough collagen. This opens up new ways to treat the disease by targeting this "memory" system.

Technical Summary

1. Problem Statement

Pancreatic Ductal Adenocarcinoma (PDAC) is characterized by a dense desmoplastic stroma rich in Type I collagen (Col-I) and chronic inflammation. While it is known that proteolytic cleavage of Col-I by matrix metalloproteases (MMPs) generates soluble fragments (specifically ¾Col-I) that activate the collagen receptor DDR1 to promote tumor growth, the mechanism by which intact collagen (iCol-I) restrains tumor growth remains unclear. Furthermore, how tumor cells integrate stromal architecture with inflammatory cues to establish adaptive, durable states (inflammatory memory) is poorly understood. Specifically, the molecular link between metabolic stress, receptor stability, and epigenetic memory in PDAC progression needs elucidation.

2. Methodology

The study employed a multidisciplinary approach combining molecular biology, biochemistry, genetics, and clinical data analysis:

• In Vitro Models: Human PDAC cell lines (PANC-1, AsPC-1, 1305, BxPC3, SW1990) and mouse PDAC cells (KPC, KC) were cultured on engineered extracellular matrices (ECM) containing either wild-type (WT) Col-I or MMP-resistant (R/R) Col-I to mimic intact vs. cleaved collagen environments. Low-glucose (LG) media was used to induce metabolic stress.
• Genetic Manipulation: CRISPR/Cas9 and shRNA were used for gene ablation/knockdown (FBXW2, DDR1, DNMTs). Overexpression constructs (WT, T519A, K663R, T519E mutants of DDR1; FBXW2; DNMTs) were utilized.
• Epigenetic Editing: A dCas9-TET2 fusion system was used to target demethylation of the FBXW2 promoter.
• Biochemical Assays: Co-immunoprecipitation (Co-IP), ubiquitination assays, in vitro kinase assays (AMPK vs. DDR1), mass spectrometry (MS) for phospho- and ubiquitylation-site mapping, and ELISA for collagen-DDR1 binding affinity.
• Structural Biology: AlphaFold3 and HDOCK were used to model the interaction between DDR1 and different collagen fragments (intact vs. ¾ fragments).
• In Vivo Models: Orthotopic transplantation of PDAC cells into Col-IWT and Col-Ir/r mice, and genetic mouse models (KrasG12D;Tp53R172H) with conditional deletion of Fbxw2 and/or Ddr1 in pancreatic epithelial cells.
• Clinical Correlation: Immunohistochemistry (IHC) on 93 human PDAC patient specimens, analysis of public databases (CPTAC-PDAC, MethMarkerDB), and survival analysis.
• Single-Cell Analysis: scATAC-seq to assess chromatin accessibility at the FBXW2 locus.

3. Key Contributions & Mechanisms

The paper identifies a metabolic-inflammatory checkpoint that controls DDR1 stability via the E3 ubiquitin ligase adaptor FBXW2.

A. The Metabolic Checkpoint (Intact Collagen $\rightarrow$ DDR1 Degradation)

1. Ligand Specificity: Intact Col-I (iCol-I) binds DDR1 with low affinity because the binding motif (GVMGFO) is buried, whereas MMP-cleaved ¾Col-I exposes this motif, allowing high-affinity binding.
2. Metabolic Stress: In the absence of high-affinity ¾Col-I, DDR1 signaling (via NRF2) is suppressed, leading to reduced macropinocytosis and mitochondrial biogenesis. This causes a drop in cellular ATP levels.
3. AMPK Activation: Low ATP activates AMPK.
4. Phosphorylation & Degradation: Activated AMPK directly phosphorylates DDR1 at Threonine 519 (T519). This phosphorylation creates a recognition site for the E3 ligase adaptor FBXW2.
5. Ubiquitination: FBXW2 recruits the CRL1 complex to ubiquitinate DDR1 (at Lysine 663), targeting it for proteasomal degradation.
6. Outcome: In restrictive stromal environments (intact collagen), this pathway degrades DDR1, suppressing tumor growth.

B. The Inflammatory Memory (Cytokines $\rightarrow$ FBXW2 Silencing)

1. Cytokine Induction: Pro-inflammatory cytokines (IL-8, IL-17A) induce the expression of DNA methyltransferases (specifically DNMT3B).
2. Epigenetic Silencing: DNMT3B mediates hypermethylation of the FBXW2 promoter, leading to transcriptional silencing of FBXW2.
3. Checkpoint Disablement: Loss of FBXW2 prevents the ubiquitination and degradation of DDR1, even in the presence of intact collagen and metabolic stress.
4. Durable Memory: This epigenetic silencing persists even after cytokine withdrawal, creating a "memory" that locks DDR1 in a stable, active state, allowing tumors to bypass stromal constraints.

4. Key Results

• FBXW2 is a Tumor Suppressor: FBXW2 knockout (KO) or knockdown (KD) in PDAC cells restores DDR1 stability and signaling, enabling robust tumor growth and liver metastasis even in Col-Ir/r mice (which lack cleaved collagen).
• AMPK-DDR1 Axis: AMPK phosphorylates DDR1 at T519. The T519A mutation (non-phosphorylatable) prevents FBXW2 binding and DDR1 degradation, rendering cells resistant to growth suppression by intact collagen.
• Clinical Correlation: In human PDAC specimens:
  • Low FBXW2 expression correlates with high DDR1, high p-NRF2, and high cleaved collagen.
  • High levels of phospho-DDR1 (pT519, the degradation signal) correlate with better patient survival and are found in early-stage tumors.
  • Low pT519 (indicating active DDR1) correlates with advanced disease and poor survival.
  • FBXW2 promoter hypermethylation is prevalent in human PDAC and inversely correlates with FBXW2 mRNA levels.
• Inflammatory Priming: Pre-treatment of cells with IL-8 or IL-17A induces FBXW2 promoter methylation. These "primed" cells exhibit sustained DDR1 signaling and aggressive tumor growth in vivo, regardless of the collagen cleavage status.
• Therapeutic Reversal: Demethylation of the FBXW2 promoter (using 5-Azacytidine or dCas9-TET2) restores FBXW2 expression, re-establishes DDR1 degradation, and suppresses tumor growth in cytokine-primed models.

5. Significance

• Mechanistic Insight: The study reveals a novel mechanism where receptor tyrosine kinase (RTK) stability is regulated not just by ligand binding, but by a metabolic checkpoint linking cellular energy status (ATP/AMPK) to receptor turnover via FBXW2.
• Inflammatory Memory: It provides a molecular basis for "inflammatory memory" in cancer, demonstrating how transient inflammatory exposure leads to permanent epigenetic reprogramming (promoter methylation) that decouples tumor growth from stromal constraints.
• Therapeutic Implications:
  • FBXW2 Restoration: Strategies to reverse FBXW2 methylation (e.g., DNMT inhibitors) could restore the natural degradation checkpoint for DDR1.
  • DDR1 Targeting: The findings suggest that DDR1 inhibition might be less effective if the degradation checkpoint is already disabled by inflammation; conversely, targeting the inflammatory-DNMT axis could sensitize tumors to stromal constraints.
  • Biomarkers: The ratio of pT519-DDR1 (degradation signal) to pY513-DDR1 (activation signal) could serve as a prognostic biomarker for PDAC progression.

In summary, the paper elucidates how PDAC cells exploit inflammatory signaling to epigenetically silence a metabolic checkpoint (FBXW2), thereby stabilizing the DDR1 receptor and enabling uncontrolled growth even in restrictive stromal environments.