Pro-domain-dependent folding and co-receptor-mediated targeting to optimize an antagonistic TGF-β monomer for gene-based delivery

This study enhances the efficacy and safety of genetically delivered TGF-β monomers for cancer immunotherapy and fibrosis treatment by incorporating a modified pro-domain to correct folding issues and fusing a CD44-binding domain to achieve over 30-fold increased inhibitory activity on target cells.

The Gist

The Big Picture: Taming a "Double-Edged Sword"

Imagine TGF-β as a very powerful, double-edged sword in your body.

• The Good Side: It helps heal wounds and keeps tissues healthy.
• The Bad Side: In diseases like cancer or fibrosis (scarring of organs), this sword goes rogue. It tells cancer cells to hide from the immune system and tells scar tissue to grow out of control.

Doctors want to stop the "Bad Side" without hurting the "Good Side." The problem is, TGF-β is like a sticky, complex puzzle piece. If you try to make a drug to block it, it often gets stuck in the wrong shape (misfolds) or causes dangerous side effects because it blocks everything, even the good stuff.

This paper is about building a better, smarter version of a "shield" that can block the bad TGF-β without causing chaos.

---

The Problem: The "Broken Puzzle Piece"

Scientists previously built a small, single-piece version of TGF-β (called mmTGF-β) that acts as a decoy. It lures the cell's receptors (the "locks") but doesn't turn the key, effectively jamming the lock so the real, dangerous TGF-β can't get in.

However, there was a manufacturing glitch:
When scientists tried to mass-produce this decoy inside mammalian cells (like a factory), the factory kept making gluey, tangled blobs instead of the neat, working decoy.

• Analogy: Imagine trying to build a complex LEGO model, but the instructions are so tricky that the pieces keep snapping together in the wrong order, creating a giant, useless ball of plastic.

The Solution 1: The "Training Wheels" (The Pro-Domain)

To fix the manufacturing glitch, the scientists looked at how nature builds TGF-β. Nature uses a "training wheels" system called a pro-domain. This is a temporary helper piece that guides the main piece into the right shape, then snaps off once the job is done.

• The Innovation: The scientists took a modified version of this "training wheels" piece and attached it to their decoy.
• The Result: The factory (the cell) suddenly started producing the decoy perfectly! The training wheels guided the folding process, ensuring the decoy came out neat and ready to work.
• The Clever Twist: They designed the connection between the training wheels and the decoy so that once the decoy is ready, a natural enzyme in the body (like a pair of scissors) snips the training wheels off, releasing the active, working decoy.

The Solution 2: The "GPS Tracker" (Targeting CD44)

Even with a perfect decoy, there's a safety risk. If you release the decoy everywhere in the body, it might block TGF-β in healthy tissues (like the heart or skin), causing side effects like heart valve issues or skin lesions.

The scientists wanted the decoy to only work on specific trouble spots: immune cells that are helping the cancer hide. These trouble cells have a specific "name tag" on their surface called CD44.

• The Innovation: They attached a GPS tracker (a specific binding domain) to their decoy. This tracker is designed to stick only to the CD44 name tag.
• The Result:
  • On Healthy Cells (No CD44): The decoy floats by, ignoring them.
  • On Trouble Cells (With CD44): The decoy grabs onto the CD44 tag and the receptor at the same time.
• The Analogy: Imagine a magnet. A normal magnet is weak and might stick to a fridge (healthy cell) or a car (cancer cell). But if you put a super-strong magnet on a specific key that only fits the car's ignition, it will only stick to the car.
• The Outcome: This "dual-locked" decoy became 30 times more powerful against the trouble cells than against healthy cells. It's like turning a sledgehammer into a laser-guided missile.

Why This Matters

1. Better Manufacturing: They solved the "gluey blob" problem, meaning they can now make enough of this drug to actually use it.
2. Safer Treatment: By adding the "GPS tracker," the drug is much less likely to hurt healthy organs. It focuses its power where it's needed most.
3. Gene Therapy Ready: This design is perfect for being delivered by viruses (like a Trojan Horse) that infect cancer cells and start churning out this super-decoy right inside the tumor.

The Bottom Line

The scientists took a promising but broken drug concept, gave it a "training wheel" to fix its shape, and added a "GPS" to make it target only the bad guys. This creates a much safer and more effective weapon against cancer and fibrosis, potentially allowing the immune system to finally win the battle.

Technical Summary

1. Problem Statement

Transforming Growth Factor-beta (TGF-β) is a potent driver of tissue fibrosis and an immunosuppressive factor in the tumor microenvironment (TME), making it a prime target for cancer immunotherapy and fibrosis treatment. However, systemic inhibition of TGF-β has historically failed in clinical trials due to severe on-target toxicities (e.g., cardiotoxicity, cutaneous squamous lesions) caused by the pleiotropic nature of TGF-β signaling.

Recent work by the authors identified an engineered mini monomer of TGF-β (mmTGF-β) that acts as a dominant-negative inhibitor. It binds TGFBR2 but cannot recruit TGFBR1, thereby blocking signaling without triggering receptor internalization. While effective when delivered via osmotic pumps or oncolytic viruses (OVs), a major bottleneck exists: inefficient folding of mmTGF-β when expressed in mammalian cells. The protein predominantly forms disulfide-linked aggregates, with very little active monomer being secreted, limiting its utility for gene-based delivery strategies (e.g., viral vectors).

2. Methodology

The authors employed a multi-pronged structural and functional engineering approach to optimize mmTGF-β for mammalian expression and targeted delivery:

• Disulfide Bond Simplification: They systematically removed specific disulfide bonds (auxiliary and cystine knot) via cysteine-to-alanine/valine substitutions to test if reducing structural complexity improved folding.
• Scaffold Swapping: They grafted the mmTGF-β "fingers" onto the scaffold of PRDC (a growth factor that folds without a pro-domain) to create chimeric constructs.
• Pro-domain Incorporation: Inspired by the natural role of the TGF-β pro-domain as an intramolecular chaperone, they engineered a fusion construct (p-mmTGF-β) containing a modified TGF-β3 pro-domain. Crucially, they deleted the "bowtie" dimerization motif (residues 227–253) to prevent pro-domain dimerization while retaining its folding assistance.
• Cleavage Optimization: They engineered furin cleavage sites (1, 2, or 3 sites) into the pro-domain to ensure efficient release of the active mmTGF-β monomer post-secretion.
• Co-receptor Targeting: To enhance specificity and safety, they fused a CD44-binding domain (derived from the parasite protein TGM1) to the C-terminus of mmTGF-β, creating a bivalent construct (mmTGF-β-ΔA-TGM1-D4/5) designed to target cells co-expressing TGFBR2 and CD44.
• Validation: Constructs were expressed in E. coli (for refolding studies) and mammalian cells (HEK293, expi293). Folding was assessed via NMR and Differential Scanning Calorimetry (DSC). Activity was measured using luciferase/mCherry reporter assays (CAGA12), Surface Plasmon Resonance (SPR), and live-cell imaging of SMAD2 nuclear accumulation.

3. Key Contributions & Results

A. Solving the Folding Problem

• Disulfide Reduction: Removing the auxiliary disulfide bond (mmTGF-β-ΔA) allowed for efficient folding and retained inhibitory activity. However, removing core cystine knot bonds led to misfolding.
• PRDC Scaffold: While PRDC-based chimeras folded well in E. coli, they showed only marginal improvement in folding and activity when expressed in mammalian cells.
• Pro-domain Solution (Breakthrough): The inclusion of a modified pro-domain (p-mmTGF-β3) dramatically improved the folding efficiency in mammalian cells.
  • The construct with a furin-cleavable site (p-mmTGF-β3-2F) resulted in the secretion of high levels of active, monomeric mmTGF-β.
  • Western blots showed that unlike the original mmTGF-β (which aggregated), the pro-domain fusion yielded a clean monomeric band that could be cleaved to release the active inhibitor.
  • Mechanism: The pro-domain acts as a chaperone during biosynthesis. Once cleaved by furin, the monomeric nature of mmTGF-β prevents the high-affinity avidity binding seen in native dimers, allowing the active inhibitor to be released and compete effectively for TGFBR2.

B. Specificity and Safety Profile

• Betaglycan Independence: The inhibitory activity of mmTGF-β was unaffected by the presence or absence of the co-receptor Betaglycan (TGFBR3), unlike native TGF-β2 which requires it for signaling.
• No BMP Cross-Reactivity: mmTGF-β did not inhibit Bone Morphogenetic Protein (BMP) signaling, confirming high specificity for the TGF-β pathway.
• No Receptor Internalization: Unlike native TGF-β, which triggers TGFBR2 internalization (a mechanism for signal attenuation), mmTGF-β binding does not induce receptor internalization. This suggests a distinct mechanism of action that avoids rapid clearance or downstream signaling activation.
• Furin Site Sufficiency: A single furin cleavage site was sufficient to release active mmTGF-β; additional sites did not significantly improve potency.

C. Co-receptor Mediated Targeting

• CD44 Fusion: The fusion of the TGM1 D4/5 domain (binding CD44) to mmTGF-β created a bivalent inhibitor.
• Avidity Effect: This construct showed a ~30-fold increase in potency against cells expressing CD44 (e.g., activated T cells, macrophages) compared to CD44-negative cells.
• Selectivity: In CD44-knockout cells, the fusion protein behaved similarly to the unmodified mmTGF-β, confirming that the enhanced potency is strictly dependent on the co-receptor interaction.

4. Significance

This study provides a critical framework for the clinical translation of gene-based TGF-β inhibition:

1. Overcoming Production Barriers: By utilizing a modified pro-domain, the authors solved the major hurdle of misfolding and aggregation in mammalian cells. This makes mmTGF-β viable for gene-based delivery (e.g., using oncolytic viruses like vaccinia), where the therapeutic protein is produced in situ by tumor cells.
2. Enhanced Safety via Targeting: The ability to fuse mmTGF-β with co-receptor binding modules (like CD44) allows for cell-type-specific inhibition. This could significantly widen the therapeutic window by concentrating the inhibitory effect on immune cells within the TME (which often overexpress CD44) while sparing healthy tissues that lack the co-receptor.
3. Mechanistic Insight: The findings clarify that the pro-domain's role in folding is separable from its role in latency, and that monomeric inhibitors can be effectively released and function without the need for complex activation mechanisms required by native latent TGF-β.

In conclusion, the optimized p-mmTGF-β constructs, particularly when combined with co-receptor targeting, represent a next-generation therapeutic strategy to suppress TGF-β-driven fibrosis and immunosuppression with improved efficacy and reduced systemic toxicity.