Targeting cancer-associated fibroblasts for treatment of ER+ breast cancer: A mathematical modeling perspective and optimization of treatment strategies

This study employs a mathematical modeling framework and optimal control theory to demonstrate that targeting cancer-associated fibroblasts through specific signaling inhibition strategies can significantly enhance the efficacy of endocrine therapies for ER+ breast cancer by mitigating estrogen-independent tumor growth.

The Gist

Imagine your body is a bustling city, and a tumor is a group of unruly squatters trying to take over a neighborhood. In ER+ breast cancer, these squatters usually need a specific key to unlock their growth: a hormone called Estrogen.

For decades, doctors have tried to treat this by removing the keys (using endocrine therapy). But sometimes, the squatters find a way to grow even without the keys. Why? Because they have hired a crew of Construction Workers living right next door.

In this paper, the scientists call these workers Cancer-Associated Fibroblasts (CAFs).

Here is the story of what the researchers discovered, explained simply:

1. The Uninvited Construction Crew (The Problem)

Normally, if you cut off the supply of estrogen (the keys), the tumor stops growing. But the researchers found that CAFs are like a super-organized construction crew.

• What they do: They don't just sit there; they actively help the tumor. They remodel the neighborhood (the tumor environment), hand out free energy, and even teach the tumor cells how to grow without needing the estrogen keys anymore.
• The Result: Even when doctors try to lock the doors (stop estrogen), the CAFs build a backdoor, allowing the cancer to keep growing. This is why some treatments stop working.

2. The Mathematical Simulator (The Tool)

Instead of just guessing in a lab, the team built a virtual video game (a mathematical model) to simulate this city.

• They programmed the rules: How fast the tumor grows, how the CAFs help, and how the hormones work.
• They tested this game against real data from mouse experiments to make sure their "game physics" matched reality.
• The Discovery: The simulation confirmed that CAFs are the reason tumors can survive when estrogen is removed. They are the "cheat codes" the cancer uses.

3. The Strategy Game (The Solution)

The researchers asked: If we can't just remove the estrogen keys, how do we stop the Construction Crew?

They tested three different "weapons" (treatments) in their simulation:

• Weapon A: Try to stop the CAFs from growing (Fire the construction crew).
• Weapon B: Stop the CAFs from helping the tumor grow (Cut the phone lines between the crew and the squatters).
• Weapon C: The standard treatment (Remove the estrogen keys).

The Big Surprise:
Using just one weapon rarely worked well.

• If you just tried to fire the crew (Weapon A), they were too tough to get rid of completely.
• If you just removed the keys (Weapon C), the crew built a backdoor anyway.

The Winning Strategy:
The simulation showed that the best approach is a "Combo Attack."
You need to use Weapon B (stop the CAFs from helping the tumor) PLUS Weapon C (keep removing the estrogen keys).

• Analogy: It's like trying to stop a house fire. You don't just turn off the gas (estrogen); you also need to cut the power lines (CAF signaling) that are keeping the fire alive. When you do both, the fire goes out much faster and stays out.

4. The Smart Scheduler (Optimization)

The researchers also asked: Do we need to use these weapons 24/7?

• Constant Treatment: Shouting "Stop!" at the tumor every single second. This is expensive and has side effects.
• Optimal Treatment: Using a smart timer. The computer calculated exactly when to hit hard and when to take a break.
• The Result: The smart schedule worked just as well as the constant shouting, but it used less "ammo" (drugs). This means patients could get the same cure with fewer side effects and lower costs.

The Takeaway

This paper tells us that to beat ER+ breast cancer, we can't just focus on the cancer cells themselves. We have to look at the neighborhood they live in.

The Construction Crew (CAFs) is the hidden boss. If we ignore them, the cancer finds a way to win. But if we combine standard hormone therapy with a new strategy that specifically targets and silences these helpers, we can shut down the tumor much more effectively.

In short: Don't just fight the bad guys; fire their accomplices, and use a smart schedule to save your energy.

Technical Summary

1. Problem Statement

Estrogen receptor-positive (ER+) breast cancer accounts for approximately 70% of breast cancer cases. While endocrine therapies targeting the estrogen receptor (ER) pathway are the standard of care, a major clinical challenge is the development of intrinsic or acquired resistance.

• The Role of CAFs: Cancer-associated fibroblasts (CAFs) within the tumor microenvironment (TME) have been identified as critical drivers of resistance. Experimental evidence (specifically from Reid et al.) suggests that CAFs can facilitate estrogen-independent tumor growth, allowing tumors to persist and proliferate even when estrogen levels are low or withdrawn.
• The Gap: There is a lack of quantitative frameworks that explicitly model the bidirectional interactions between tumor cells, hormonal signaling, and CAFs to optimize treatment strategies. Current models often treat CAFs as passive bystanders or do not incorporate them into optimal control problems for therapy scheduling.

2. Methodology

The authors developed a mechanistic mathematical framework based on systems of nonlinear ordinary differential equations (ODEs) and applied optimal control theory to evaluate therapeutic strategies.

A. Mathematical Modeling

The study progresses through three modeling stages:

1. Baseline Model (No CAFs): A system describing tumor volume ($T$), ER$\alpha$ concentration ($ER$), and estrogen-bound ER complexes ($E2ER$). It incorporates logistic growth, Michaelis-Menten kinetics for $E2ER$-mediated proliferation, and binding/unbinding kinetics.
2. Extended Model (With CAFs): The model was extended to include CAF population dynamics ($CAF$). Key biological mechanisms incorporated:
   • Bidirectional Signaling: CAFs promote tumor growth (proliferative effect), and tumor cells trigger CAF growth.
   • ER Inhibition: CAFs inhibit the translation of ER$\alpha$, reducing the tumor's sensitivity to endocrine therapy.
   • Estrogen Independence: The model allows tumor growth to be driven by CAFs even when estrogen is depleted.
3. Structural Identifiability: Using the GenSSI 2.0 toolbox, the authors proved that the extended model is structurally globally identifiable, ensuring that model parameters can be uniquely determined from the available data (tumor volume and ER intensity).

B. Parameter Estimation

• Data Source: Experimental data from Reid et al. involving MCF7 cells (ER+ breast cancer) and human CAFs orthotopically transplanted into NSG mice.
• Conditions: Three estrogen (E2) dosing regimens were simulated: High (0.5 mg), Medium (0.1 mg), and Low (0.025 mg).
• Calibration: Parameters were inferred using Monolix 2024R1. Random effects were applied to account for inter-mouse variability. The model successfully reproduced the experimental observation that tumors co-transplanted with CAFs continued to grow after estrogen withdrawal, whereas those without CAFs regressed.

C. Treatment Modeling & Optimization

The authors formulated an Optimal Control Problem (OCP) to minimize tumor volume while penalizing treatment costs (side effects).

• Control Variables: Three time-dependent control functions ($u_1, u_2, u_3$) representing:
  • Type I ($u_1$): Inhibition of CAF proliferation.
  • Type II ($u_2$): Inhibition of the proliferative effect of CAFs on tumor cells.
  • Type III ($u_3$): Inhibition of E2-ER binding (Endocrine therapy).
• Objective Function: Minimize $\int (T + \text{cost of controls}) dt$.
• Solution Method: The Pontryagin's Maximum Principle was used to derive the adjoint system and optimal control laws. The system was solved numerically using a forward-backward sweep method.

3. Key Contributions

1. Mechanistic Integration: The paper provides one of the first ODE-based models to explicitly quantify how CAFs reprogram ER signaling and drive estrogen-independent growth in ER+ breast cancer.
2. Optimization Framework: It introduces a rigorous optimal control framework to determine when and how strongly to apply CAF-targeted therapies versus endocrine therapies, moving beyond static "constant" dosing.
3. Identification of Synergistic Strategies: The study mathematically demonstrates that monotherapies (targeting only CAFs or only hormones) are often insufficient, whereas specific combinations yield superior results.

4. Key Results

A. Constant Treatment Simulations

• High Estrogen (0.5 mg): The most effective strategy was the combination of Type II (blocking CAF-to-tumor signaling) and Type III (endocrine therapy). Type I (blocking CAF growth alone) was the least effective.
• Low Estrogen (0.025 mg): CAFs became the dominant driver of growth. The combination of Type II and Type III remained the most effective, significantly outperforming monotherapies.
• Non-Linear Effects: The study observed a non-intuitive non-linear relationship between the initial CAF/Tumor ratio and treatment efficacy. Intermediate CAF populations sometimes resulted in larger tumors than very high or very low populations due to complex feedback loops.

B. Optimal Control Results

• Dynamic Scheduling: Optimal control strategies revealed that continuous maximal inhibition is not necessary. The optimal schedules often involved periods of inactivity or reduced dosing, suggesting that "adaptive" or "intermittent" dosing could reduce toxicity without compromising efficacy.
• Superiority over Constant Dosing: For most scenarios, the optimal time-varying control strategies achieved better tumor reduction with lower total drug usage compared to constant high-dose (99% inhibition) regimens.
• Specific Findings by Estrogen Level:
  • High/Medium E2: Combination of Type II + Type III was optimal.
  • Low E2: The optimal strategy required longer durations of Type II and Type III application to counteract the strong CAF-driven resistance.

C. Quantitative Outcomes

• Tumor Reduction: The combination of Type II and Type III achieved the highest percentage reduction in final tumor volume (up to ~93% in high E2 conditions with 99% inhibition).
• Cost Function: In low estrogen conditions, optimal treatment minimized the cost functional ($J$) primarily by reducing the treatment cost component ($J_2$) while maintaining low tumor burden, proving the efficiency of dynamic scheduling.

5. Significance and Implications

• Therapeutic Strategy: The study provides strong theoretical evidence that targeting CAFs is essential for overcoming endocrine resistance in ER+ breast cancer. Specifically, blocking the signaling from CAFs to tumor cells (Type II) is more critical than simply killing CAFs (Type I).
• Clinical Translation: The findings support the development of combination therapies that pair standard endocrine drugs with agents that disrupt CAF-tumor crosstalk (e.g., inhibitors of TGF-$\beta$ or SDF-1/CXCR4 pathways).
• Personalized Medicine: The model suggests that treatment protocols should be tailored based on the initial CAF burden and estrogen levels. The "optimal control" approach offers a blueprint for digital twin applications, where patient-specific models could be used to design adaptive dosing schedules that maximize efficacy and minimize toxicity.
• Future Directions: The authors note that future work should address CAF heterogeneity (different subtypes) and validate these pharmacodynamic parameters in clinical datasets to further refine the model for human application.

In conclusion, this paper bridges the gap between experimental observations of CAF-mediated resistance and clinical treatment design, demonstrating that mathematical optimization can identify superior, less toxic, and more effective treatment schedules for ER+ breast cancer.