armored-cart-design-agent

<!--
# COPYRIGHT NOTICE
# This file is part of the "Universal Biomedical Skills" project.
# Copyright (c) 2026 MD BABU MIA, PhD <md.babu.mia@mssm.edu>
# All Rights Reserved.
#
# This code is proprietary and confidential.
# Unauthorized copying of this file, via any medium is strictly prohibited.
#
# Provenance: Authenticated by MD BABU MIA

-->

---
name: 'armored-cart-design-agent'
description: 'AI-powered design of armored CAR-T cells with cytokine/chemokine expression for enhanced solid tumor efficacy, including IL-12, IL-15, IL-18, and IL-7 armoring strategies.'
measurable_outcome: Execute skill workflow successfully with valid output within 15 minutes.
allowed-tools:
  - read_file
  - run_shell_command
---


# Armored CAR-T Design Agent

The **Armored CAR-T Design Agent** provides AI-assisted design of next-generation armored CAR-T cells engineered to express cytokines, chemokines, or other enhancing factors. These armored T cells overcome solid tumor challenges including immunosuppressive TME, poor trafficking, and T cell exhaustion, with recent clinical success in lymphoma (IL-18) and ongoing trials with IL-12, IL-15, and IL-7.

## When to Use This Skill

* When designing CAR-T cells for solid tumor applications.
* For selecting optimal armoring payloads (cytokines, chemokines).
* To optimize cytokine expression levels and regulation.
* When engineering safety switches for armored constructs.
* For predicting armored CAR-T efficacy and safety profiles.

## Core Capabilities

1. **Armoring Payload Selection**: Choose optimal cytokines for tumor type.

2. **Expression Level Optimization**: Balance efficacy vs toxicity.

3. **Inducible System Design**: Engineer regulated expression systems.

4. **Safety Switch Integration**: Design kill switches and controls.

5. **Construct Optimization**: Optimize transgene configuration.

6. **Efficacy Prediction**: Predict enhanced tumor killing.

## Armoring Strategies

| Cytokine | Mechanism | Clinical Status | Tumor Types |
|----------|-----------|-----------------|-------------|
| IL-12 | Th1 polarization, IFN-gamma | Phase I/II | Solid tumors |
| IL-15 | T/NK persistence | Phase I/II | Hematologic, solid |
| IL-18 | Inflammasome, IFN-gamma | Phase I (promising) | Lymphoma |
| IL-7 | T cell survival | Phase I | Multiple |
| IL-21 | T cell proliferation | Preclinical | Multiple |
| CCL19/21 | T cell trafficking | Preclinical | Solid tumors |

## Construct Architecture Options

| Component | Options | Consideration |
|-----------|---------|---------------|
| Promoter | EF1a, PGK, CAG, NFAT-inducible | Expression level/timing |
| Signal Peptide | Native, IL-2ss, IgK | Secretion efficiency |
| Cytokine | Membrane-bound vs secreted | Local vs systemic |
| Linker | T2A, P2A, IRES | Co-expression efficiency |
| Kill Switch | iCasp9, HSV-TK, CD20 | Safety control |
| Position | Before/after CAR | Expression balance |

## Workflow

1. **Input**: Target tumor type, TME characteristics, CAR design.

2. **Payload Selection**: Rank armoring strategies for tumor context.

3. **Expression Design**: Optimize promoter, levels, regulation.

4. **Safety Engineering**: Add appropriate control switches.

5. **Construct Assembly**: Generate optimized DNA sequence.

6. **Efficacy Prediction**: Model enhanced killing and persistence.

7. **Output**: Optimized armored CAR construct with annotations.

## Example Usage

**User**: "Design an armored CAR-T for pancreatic cancer targeting mesothelin with IL-12 armoring for TME remodeling."

**Agent Action**:
```bash
python3 Skills/Immunology_Vaccines/Armored_CART_Design_Agent/design_armored_cart.py \
    --car_target mesothelin \
    --tumor_type pancreatic \
    --armoring_payload IL-12 \
    --expression_system NFAT_inducible \
    --safety_switch iCasp9 \
    --backbone lentiviral \
    --optimize_codon human \
    --output armored_cart_design/
```

## Output Components

| Output | Description | Format |
|--------|-------------|--------|
| Construct Sequence | Full transgene DNA | .fasta, .gb |
| Construct Map | Annotated visualization | .png, .pdf |
| Expression Model | Predicted levels | .json |
| Safety Analysis | Risk assessment | .json |
| Manufacturing Guide | Production recommendations | .md |
| Predicted Efficacy | Tumor killing model | .json |

## IL-12 Armoring Details

| Aspect | Design Choice | Rationale |
|--------|---------------|-----------|
| Configuration | Tethered IL-12 (p70) | Localized, reduced toxicity |
| Expression | NFAT-inducible | Activation-dependent |
| Dose | Low-level expression | Safety optimization |
| Combination | With PD-1 knockout | Enhanced activity |

## IL-18 Armoring Details

| Aspect | Design Choice | Rationale |
|--------|---------------|-----------|
| Configuration | Secreted mature IL-18 | Enhanced IFN-gamma |
| Expression | Constitutive or inducible | Context-dependent |
| Clinical Results | Lymphoma responses | Validated approach |
| Combination | With IL-21 | Synergistic |

## IL-15 Armoring Details

| Aspect | Design Choice | Rationale |
|--------|---------------|-----------|
| Configuration | Membrane-tethered IL-15/IL-15Ra | Cis-presentation |
| Expression | Constitutive moderate | Persistence without toxicity |
| Benefit | Reduced IL-2 dependence | Manufacturing advantage |
| Safety | Lower CRS risk | Clinical benefit |

## AI/ML Components

**Payload Selection**:
- TME profiling to match cytokine needs
- Multi-objective optimization
- Clinical outcome modeling

**Expression Optimization**:
- Promoter strength prediction
- Codon optimization
- mRNA stability modeling

**Safety Prediction**:
- CRS/ICANS risk modeling
- Off-tumor activity prediction
- Systemic cytokine levels

## Safety Considerations

| Risk | Mitigation | Implementation |
|------|------------|----------------|
| Cytokine storm | Inducible expression | NFAT promoter |
| Systemic toxicity | Membrane tethering | Localized effect |
| Uncontrolled proliferation | Kill switch | iCasp9 |
| On-target off-tumor | Regulatable CAR | Logic gates |

## Clinical Trials (2025-2026)

| Trial | Armoring | Target | Cancer | Status |
|-------|----------|--------|--------|--------|
| NCT03721068 | IL-18 | CD19 | Lymphoma | Phase I (positive) |
| NCT04119024 | IL-12 | GD2 | Neuroblastoma | Phase I |
| NCT03932565 | IL-15/21 | CD19 | B-ALL | Phase I |
| Multiple | IL-7/CCL19 | Various | Solid | Preclinical |

## Prerequisites

* Python 3.10+
* Biopython for sequence handling
* CAR design databases
* Codon optimization tools
* Structure prediction (optional)

## Related Skills

* CART_Design_Optimizer_Agent - Base CAR optimization
* NK_Cell_Therapy_Agent - NK cell engineering
* Cytokine_Storm_Analysis_Agent - Safety analysis
* TCell_Exhaustion_Analysis_Agent - Exhaustion prevention

## Manufacturing Considerations

| Aspect | Armored CAR Challenge | Solution |
|--------|----------------------|----------|
| Vector Size | Larger transgene | Optimize construct |
| Transduction | Lower efficiency | Increase MOI |
| Expansion | Cytokine effects | Tune expression |
| Characterization | Complex phenotype | Enhanced QC |

## Special Considerations

1. **Tumor Type Matching**: Different tumors need different armoring
2. **Expression Timing**: Constitutive vs inducible tradeoffs
3. **Dose Finding**: Balance efficacy vs toxicity
4. **Combination**: Consider with checkpoint knockout
5. **Manufacturing**: Larger constructs affect production

## Efficacy Enhancement Mechanisms

| Mechanism | Cytokine | Effect |
|-----------|----------|--------|
| Persistence | IL-15, IL-7 | Longer survival |
| TME Remodeling | IL-12 | M2→M1, DC activation |
| Bystander Killing | IL-18 | Enhanced IFN-gamma |
| Trafficking | CCL19/21 | T cell recruitment |
| Anti-exhaustion | IL-21 | Stem-like maintenance |

## Author

AI Group - Biomedical AI Platform


<!-- AUTHOR_SIGNATURE: 9a7f3c2e-MD-BABU-MIA-2026-MSSM-SECURE -->