# Fc-Engineered Anti-FAP Antibodies Enhance NK Cell-Mediated ADCC and Remodel the Immunosuppressive Tumor Microenvironment

## Metadata

- **SPORE ID**: SPR-2026-FBCA
- **Domaines**: Monoclonal and Polyclonal Antibodies Research x Peptidase Inhibition and Analysis
- **Date de generation**: 2026-08-11
- **Panel consensus score**: 6.5/10
- **Novelty score**: 0.8/1.0
- **Panel verdict**: publish_brief

## Abstract

If an anti-FAP antibody is engineered with Fc mutations (S239D/I332E) to increase FcγRIIIa affinity, then FAP+ CAFs will be depleted via NK cell-mediated ADCC, leading to a measurable reduction in immunosuppressive cytokines and increased CD8+ T cell infiltration, because the enhanced Fc-FcγRIIIa interaction overcomes the low baseline ADCC activity against stromal targets.

The proposed mechanism involves 6 causal steps: (1) The Fc-engineered anti-FAP antibody (e.g., S239D/I332E) binds to FAP antigen on  -> (2) The engineered Fc domain binds to FcγRIIIa (CD16a) on NK cells with increased af -> (3) This cross-linking triggers NK cell degranulation, releasing perforin and granzy -> (4) The depletion of FAP+ CAFs reduces the secretion of immunosuppressive cytokines .

Literature grounding on 14 verified references yields a novelty score of 0.8 (novel). A 3-phase experimental protocol (budget: €22k-117k, timeline: 10-16 months) is proposed, starting with in silico validation. A panel of 5 expert reviewers reached a consensus score of 6.5/10.

## 1. Hypothese et mecanisme propose

### 1.1 Formulation formelle

If an anti-FAP antibody is engineered with Fc mutations (S239D/I332E) to increase FcγRIIIa affinity, then FAP+ CAFs will be depleted via NK cell-mediated ADCC, leading to a measurable reduction in immunosuppressive cytokines and increased CD8+ T cell infiltration, because the enhanced Fc-FcγRIIIa interaction overcomes the low baseline ADCC activity against stromal targets.

### 1.2 Variables

**Variables independantes :**

| Variable | Type | Plage | Unite |
|----------|------|-------|-------|
| Fc engineering variant of anti-FAP antibody | categorical | WT (FcγRIIIa Kd ≈ 500 nM), S239D/I332E (Kd ≈ 50 nM), afucosylated (Kd ≈ 100 nM) | nM (binding affinity) |
| FcγRIIIa (CD16a) polymorphism status of effector NK cells | categorical | V/V (high affinity), V/F (intermediate), F/F (low affinity) | genotype |
| Antibody concentration | continuous | 0.1 - 100 μg/mL | μg/mL |

**Variables dependantes :**

| Variable | Type | Direction attendue | Unite |
|----------|------|-------------------|-------|
| Specific lysis of FAP+ CAFs (ADCC) | continuous | increase | % lysis (0-100%) |
| Tumor interstitial concentration of CCL2 and WNT2 | continuous | decrease | pg/mg total protein |
| Intratumoral CD8+ T cell density | continuous | increase | cells/mm² |
| Intratumoral Treg (CD4+CD25+FoxP3+) density | continuous | decrease | cells/mm² |
| Tumor volume (in vivo) | continuous | decrease | mm³ |

### 1.3 Chaine causale

1. Step 1: The Fc-engineered anti-FAP antibody (e.g., S239D/I332E) binds to FAP antigen on CAFs with high affinity (Kd < 10 nM) via its Fab domain.
1. Step 2: The engineered Fc domain binds to FcγRIIIa (CD16a) on NK cells with increased affinity (Kd < 100 nM), overcoming the inhibitory signals from HLA-E on CAFs.
1. Step 3: This cross-linking triggers NK cell degranulation, releasing perforin and granzyme B, leading to apoptosis of FAP+ CAFs within 4-24 hours.
1. Step 4: The depletion of FAP+ CAFs reduces the secretion of immunosuppressive cytokines (WNT2, CCL2) by >50%.
1. Step 5: Reduced CCL2 levels decrease Treg recruitment, and reduced WNT2 signaling decreases T cell exhaustion, leading to increased CD8+ T cell infiltration and function.
1. Step 6: The remodeled TME (increased CD8/Treg ratio) enhances the efficacy of checkpoint inhibitors like anti-PD-1.

**Hypotheses cles :**

- FAP is expressed on the cell surface of CAFs at sufficient density (>10,000 copies/cell) for efficient ADCC.
- NK cells are present in the tumor microenvironment or can be recruited to the tumor site.
- The Fc-engineered antibody does not bind to off-target cells expressing FAP (e.g., bone marrow stromal cells) causing systemic toxicity.
- The immunosuppressive phenotype of CAFs is reversible upon their depletion, not compensated by other stromal cell types.

**Inconnues identifiees :**

- We do not know the threshold of FAP expression density required for efficient NK cell-mediated ADCC.
- We do not know whether the in vitro ADCC activity will translate to in vivo efficacy given the hypoxic and metabolically challenging TME.
- We do not know the extent to which CAF depletion will lead to compensatory mechanisms from other fibroblast subsets.

### 1.4 Conditions aux limites

- **FAP expression density on CAFs must be above a threshold of approximately 10,000 receptors per cell.** — Below this density, the avidity of the antibody-FAP interaction may be insufficient to trigger stable immunological synapse formation and NK cell activation, as shown in other ADCC models.
- **The tumor microenvironment must contain or be able to recruit NK cells (CD3-CD56+).** — The proposed mechanism relies entirely on NK cell-mediated ADCC. In 'cold' tumors with no NK cell infiltration, the antibody will be ineffective.
- **The treatment is limited to tumors where FAP is expressed on CAFs and not on malignant cells.** — If FAP is expressed on the tumor cells themselves, the antibody may directly kill cancer cells, confounding the interpretation of the CAF-depletion mechanism.
- **The Fc-engineered antibody must not induce significant systemic toxicity due to FAP expression in healthy tissues (e.g., bone marrow).** — FAP is expressed in some normal adult tissues. A therapeutic window must exist where tumor-associated CAFs are preferentially targeted.

### 1.5 Cadre theorique

Antibody-dependent cell-mediated cytotoxicity (ADCC) and the cancer-immunity cycle, specifically the concept of targeting the immunosuppressive tumor stroma to relieve immune suppression and restore anti-tumor immunity.

## 2. Etat de l'art et positionnement

### 2.1 Travaux les plus proches

- **[2018] An Oncolytic Virus Expressing a T-cell Engager Simultaneously Targets Cancer and Immunosuppressive Stromal Cells.** — [10.1158/0008-5472.CAN-18-1750](https://doi.org/10.1158/0008-5472.CAN-18-1750)
  - Similarite: related
  - Difference cle: This paper uses a T-cell engager (BiTE) format targeting FAP and CD3, engaging T cells rather than NK cells. The SPORE hypothesis specifically proposes using Fc-engineered antibodies to engage NK cells via ADCC, a distinct mechanism and cell type.
- **[2024] Mesothelin CAR T-cells secreting anti-FAP/anti-CD3 molecules efficiently target pancreatic adenocarcinoma and its stroma** — [10.1158/1078-0432.CCR-23-3841](https://doi.org/10.1158/1078-0432.CCR-23-3841)
  - Similarite: related
  - Difference cle: This work uses CAR-T cells secreting bispecific T-cell engagers to target FAP+ CAFs. The SPORE hypothesis proposes a different approach: using Fc-engineered antibodies to recruit NK cells, not T cells, for ADCC against FAP+ CAFs.
- **[2021] Fibroblast Activation Protein (FAP)-Targeted CAR-T Cells: Launching an Attack on Tumor Stroma** — [10.2147/ITT.S291767](https://doi.org/10.2147/ITT.S291767)
  - Similarite: related
  - Difference cle: This is a review of CAR-T cell therapy targeting FAP. The SPORE hypothesis is distinct as it focuses on Fc-engineered antibodies and NK cell-mediated ADCC, not CAR-T cells.

### 2.2 Base de preuves

- **[2018] Differential control of human Treg and effector T cells in tumor immunity by Fc-engineered anti–CTLA-4 antibody** — [10.1073/pnas.1812186116](https://doi.org/10.1073/pnas.1812186116)
  - Type: analogous | Citations: 178
  - Provides a direct precedent for Fc-engineering of an antibody to enhance its cytotoxic effector function (ADCC) against an immunosuppressive cell population (Tregs) in the tumor microenvironment. This is analogous to the proposed strategy of Fc-engineering anti-FAP antibodies to deplete immunosuppressive CAFs.
- **[2018] Antibodies to Costimulatory Receptor 4‐1BB Enhance Anti‐tumor Immunity via T Regulatory Cell Depletion and Promotion of CD8 T Cell Effector Function** — [10.1016/j.immuni.2018.09.014](https://doi.org/10.1016/j.immuni.2018.09.014)
  - Type: analogous | Citations: 141
  - Demonstrates that the anti-tumor efficacy of an antibody can be enhanced by engineering its Fc region to deplete immunosuppressive cells (Tregs) via FcγR-mediated mechanisms. This supports the core concept of the SPORE hypothesis: Fc engineering can be used to turn an antibody into a tool for eliminating immunosuppressive cells.
- **[2021] Engineered exosome-like nanovesicles suppress tumor growth by reprogramming tumor microenvironment and promoting tumor ferroptosis.** — [10.1016/j.actbio.2021.09.003](https://doi.org/10.1016/j.actbio.2021.09.003)
  - Type: indirect | Citations: 176
  - Shows that targeting FAP+ CAFs can reduce immunosuppressive cells (Tregs, MDSCs, M2 macrophages) and increase effector T cell infiltration, validating FAP+ CAFs as a therapeutic target to remodel the TME. This supports the downstream goal of the SPORE hypothesis.
- **[2025] FAP upregulates PD-L1 expression in cancer-associated fibroblasts to exacerbate T cells dysfunction and suppress anti-tumor immunity.** — [10.1016/j.canlet.2025.217475](https://doi.org/10.1016/j.canlet.2025.217475)
  - Type: direct | Citations: 37
  - Directly demonstrates that FAP+ CAFs are immunosuppressive by upregulating PD-L1, which leads to T cell dysfunction. This provides a strong mechanistic rationale for eliminating FAP+ CAFs, as proposed in the SPORE hypothesis, to relieve immunosuppression.
- **[2024] Investigating the complex interplay between fibroblast activation protein α-positive cancer associated fibroblasts and the tumor microenvironment in the context of cancer immunotherapy** — [10.3389/fimmu.2024.1352632](https://doi.org/10.3389/fimmu.2024.1352632)
  - Type: direct | Citations: 18
  - Provides evidence that FAP+ CAFs are associated with immunosuppressive circuits that limit anti-tumor immunity, supporting the hypothesis that targeting these cells could remodel the TME.
- **[2025] Fibroblast activation protein and the tumour microenvironment: challenges and therapeutic opportunities** — [10.3389/or.2025.1617487](https://doi.org/10.3389/or.2025.1617487)
  - Type: indirect | Citations: 27
  - This review highlights the role of FAP+ CAFs in creating an immunosuppressive TME and discusses challenges in targeting FAP, providing context for the proposed approach.
- **[2021] Novel TLR 7/8 agonists for improving NK cell mediated antibody-dependent cellular cytotoxicity (ADCC)** — [10.1038/s41598-021-83005-6](https://doi.org/10.1038/s41598-021-83005-6)
  - Type: analogous | Citations: 27
  - Demonstrates that enhancing NK cell-mediated ADCC can improve the anti-cancer efficacy of monoclonal antibodies. This supports the central mechanism of the SPORE hypothesis, which relies on NK cell-mediated ADCC.
- **[2022] High-affinity FcγRIIIa genetic variants and potent NK cell-mediated antibody-dependent cellular cytotoxicity (ADCC) responses contributing to severe COVID-19** — [10.1016/j.gim.2022.04.005](https://doi.org/10.1016/j.gim.2022.04.005)
  - Type: indirect | Citations: 21
  - Shows that genetic variants in FcγRIIIa (CD16a) influence the potency of NK cell-mediated ADCC. This supports the hypothesis's prediction that FcγR polymorphisms could lead to differential patient responses to an Fc-engineered anti-FAP antibody.
- **[2021] ADCC-Inducing Antibody Trastuzumab and Selection of KIR-HLA Ligand Mismatched Donors Enhance the NK Cell Anti-Breast Cancer Response** — [10.3390/cancers13133232](https://doi.org/10.3390/cancers13133232)
  - Type: indirect | Citations: 18
  - Demonstrates that NK cell-mediated ADCC is a key mechanism of action for a therapeutic antibody (trastuzumab) and that NK cell activity can be enhanced. This supports the feasibility of relying on NK cell ADCC for the proposed anti-FAP antibody.

### 2.3 Contre-preuves et limitations connues

- **[minor] Design of a Fibroblast Activation Protein–Targeted Radiopharmaceutical Therapy with High Tumor–to–Healthy-Tissue Ratios**
  - This paper focuses on using FAP as a target for radiopharmaceutical therapy, not for antibody-mediated cell killing. It does not directly contradict the hypothesis but highlights an alternative approach to targeting FAP, suggesting that the field is exploring various modalities.
- **[minor] Alteration of the tumor stroma using a consensus DNA vaccine targeting Fibroblast Activation Protein (FAP) synergizes with anti-tumor vaccine therapy in mice**
  - This paper uses a DNA vaccine to target FAP, which induces a T-cell mediated immune response, not an NK cell-mediated ADCC response. It does not contradict the hypothesis but shows an alternative immune-based strategy for targeting FAP.

### 2.4 Evaluation de nouveaute

- **Score**: 0.8/1.0
- **Verdict**: novel

## 3. Predictions falsifiables

| # | Prediction | Borne quantitative | Methode | H0 | Test statistique |
|---|-----------|-------------------|---------|-----|-----------------|
| 1 | The S239D/I332E Fc-engineered anti-FAP antibody will induce higher ADCC against FAP+ CAFs compared to the wild-type antibody. | At an antibody concentration of 10 μg/mL and an E:T ratio of 10:1, the S239D/I332E variant will induce ≥ 40% specific lysis of FAP+ CAFs, compared to < 10% for the WT antibody, over a 4-hour chromium-51 release assay. | Standard 4-hour 51Cr-release assay or LDH release assay using primary human NK cells (isolated from healthy donors) and FAP+ CAFs (isolated from tumor samples or differentiated from fibroblasts). | H0: There is no significant difference in specific lysis between the S239D/I332E variant and the WT antibody (mean difference = 0). | Two-way ANOVA with antibody variant and concentration as factors, followed by Tukey's HSD post-hoc test, alpha=0.05. |
| 2 | NK cells from donors with the high-affinity FcγRIIIa V/V genotype will mediate higher ADCC with the Fc-engineered antibody than NK cells from F/F donors. | At an antibody concentration of 1 μg/mL and an E:T ratio of 10:1, NK cells from V/V donors will show a ≥ 2-fold increase in specific lysis of FAP+ CAFs compared to NK cells from F/F donors. | Genotyping of FcγRIIIa (rs396991) by PCR-RFLP or TaqMan assay. ADCC assay as described above, comparing NK cells from V/V (n=5) and F/F (n=5) donors. | H0: There is no significant difference in specific lysis between V/V and F/F donor NK cells. | Unpaired two-tailed t-test, alpha=0.05. |
| 3 | In a xenograft model, treatment with the Fc-engineered anti-FAP antibody will reduce tumor volume and increase CD8+ T cell infiltration compared to control. | In a subcutaneous MC38 or HT-29 xenograft model with co-injected FAP+ CAFs, treatment with 10 mg/kg of the Fc-engineered antibody (twice weekly for 3 weeks) will result in a ≥ 50% reduction in tumor volume (mm³) and a ≥ 2-fold increase in intratumoral CD8+ T cells (cells/mm²) compared to the vehicle control group. | Tumor volume measured by caliper. Tumor-infiltrating lymphocytes analyzed by flow cytometry and immunohistochemistry (IHC) for CD8 and FoxP3. | H0: There is no significant difference in tumor volume or CD8+ T cell density between the Fc-engineered antibody group and the vehicle control group. | One-way ANOVA with treatment as a factor, followed by Dunnett's post-hoc test, alpha=0.05. |

## 4. Protocole experimental

**Timeline globale**: 10-16 months
**Budget global**: €22k-117k

### 4.1 Phase 1 — In Silico Validation & Biomarker Stratification

**Objectif**: To computationally validate the mechanistic feasibility of the hypothesis, identify potential off-target risks, and stratify patient/donor populations most likely to respond, thereby de-risking subsequent experimental phases.

**Methodologie**: 1. **Structural Modeling & Binding Analysis:** Use high-resolution structures of FAP (PDB: 1Z68) and FcγRIIIa (PDB: 3AY4) to perform in silico mutagenesis and molecular docking. Tools: MOE (Molecular Operating Environment) or Schrödinger Suite. Simulate the S239D/I332E mutations and afucosylation to confirm enhanced FcγRIIIa binding affinity (ΔΔG calculations) and predict the stability of the Fc-FcγRIIIa complex.
2. **Transcriptomic Deconvolution & Threshold Analysis:** Analyze publicly available single-cell RNA-seq datasets of solid tumors (e.g., from the Human Cell Atlas, TCGA, or GEO) to quantify FAP expression on CAFs (average copies per cell) and NK cell infiltration (CD3-CD56+ signature). Use tools like CIBERSORTx or Scissor to deconvolve bulk RNA-seq data and estimate the proportion of FAP+ CAFs and NK cells. Determine if the FAP expression threshold (>10,000 copies/cell) is met in a significant subset of patients.
3. **ADCC Potency Prediction Model:** Develop a simple mathematical model (e.g., using Python with libraries like SciPy) based on the law of mass action to simulate the ADCC response. Inputs: antibody concentration (0.1-100 µg/mL), FcγRIIIa affinity (WT vs. S239D/I332E), FAP density, and NK cell avidity. Output: predicted % specific lysis. Calibrate the model using published data on ADCC against other targets (e.g., trastuzumab/Her2).
4. **Off-Target Toxicity Risk Assessment:** Query large-scale proteomic and transcriptomic databases (e.g., Human Protein Atlas, GTEx) to map FAP expression in healthy adult tissues, specifically bone marrow stromal cells and adipose tissue. Use this to estimate a potential therapeutic window.

- Cout: €1,500-3,000 (mainly for software licenses and HPC time)
- Duree: 4-6 weeks
- Equipement: High-performance computing workstation (e.g., 32-core CPU, 128GB RAM)
- Logiciels: Schrödinger Suite (or MOE), Python 3.x with NumPy/SciPy, R with Seurat/CIBERSORTx, PyMOL for visualization

**Criteres de succes :**

- Predicted ADCC improvement of S239D/I332E over WT: ≥ 4-fold increase in predicted % specific lysis at 10 µg/mL and E:T 10:1
- FAP expression threshold: ≥ 30% of analyzed tumor samples (e.g., from TCGA) show FAP+ CAFs with estimated density > 10,000 copies/cell
- NK cell infiltration: ≥ 20% of analyzed tumor samples show a significant NK cell signature (top quartile of CD3-CD56+ score)
- Therapeutic index estimation: Predicted tumor:FAP expression ratio ≥ 10:1 compared to bone marrow stromal cells

- **GO**: All four success criteria are met: predicted ADCC improvement is ≥4-fold, FAP density and NK infiltration thresholds are met in a substantial patient subset, and the therapeutic index is acceptable.
- **NO-GO**: The predicted ADCC improvement is <2-fold, OR FAP density/NK infiltration thresholds are met in <10% of samples, OR the therapeutic index is <5:1, suggesting a high risk of systemic toxicity.
- **PIVOT**: If the ADCC model predicts that the afucosylated variant is significantly more potent than S239D/I332E, pivot to prioritize the afucosylated variant for experimental validation. If NK infiltration is universally low, pivot to consider a bispecific antibody (e.g., anti-FAP x anti-CD16) to recruit NK cells more effectively.

### 4.2 Phase 2 — Minimal Experimental Validation (In Vitro ADCC)

**Objectif**: To empirically confirm the central mechanism of the hypothesis: that the S239D/I332E Fc-engineered anti-FAP antibody mediates significantly higher NK cell ADCC against FAP+ CAFs compared to the WT antibody, and that this effect is dependent on FcγRIIIa affinity.

**Methodologie**: 1. **Cell Line Generation & Characterization:**
   - Obtain primary human fibroblasts (e.g., from ATCC or PromoCell) and differentiate them into CAF-like cells using TGF-β (10 ng/mL) for 72 hours. Confirm FAP expression by flow cytometry (anti-FAP antibody, e.g., FAP-PE, R&D Systems) and quantify receptor density using QuantiBRITE PE beads (BD Biosciences) to ensure >10,000 copies/cell.
   - Isolate primary NK cells from healthy donor buffy coats (e.g., from a local blood bank) using negative selection (Miltenyi NK Cell Isolation Kit). Genotype donors for FcγRIIIa (rs396991) using TaqMan SNP Genotyping Assay (Thermo Fisher). Select 3 V/V and 3 F/F donors.
2. **ADCC Assay (LDH Release):**
   - Use the CytoTox 96 Non-Radioactive Cytotoxicity Assay (Promega).
   - Co-culture CAFs (targets) with NK cells (effectors) at E:T ratios of 1:1, 5:1, and 10:1.
   - Test antibodies: WT anti-FAP (e.g., FAP MAb, R&D Systems) and S239D/I332E variant (produced via site-directed mutagenesis and recombinant expression in Expi293F cells, or purchased from a commercial vendor like GenScript).
   - Test antibody concentrations: 0.1, 1, 10, and 100 µg/mL.
   - Incubate for 4 hours at 37°C. Measure LDH release per manufacturer's protocol.
   - Calculate % specific lysis = (Experimental - Effector Spontaneous - Target Spontaneous) / (Target Maximum - Target Spontaneous) x 100.
3. **Experimental Controls:**
   - Include a no-antibody control (NK cells + CAFs only) to measure background lysis.
   - Include an isotype control antibody (e.g., human IgG1) to confirm specificity.
   - Include a positive control (e.g., anti-EGFR antibody Cetuximab with EGFR+ target cells) to validate NK cell functionality.

- Cout: €8,000-15,000 (mainly for primary cells, recombinant antibodies, and kits)
- Duree: 2-3 months
- Equipement: CO2 incubator, Biosafety cabinet, Flow cytometer (e.g., BD Accuri C6), Plate reader for absorbance (e.g., Tecan Infinite M200), Centrifuge, Hemocytometer or automated cell counter
- Logiciels: FlowJo for FACS data analysis, GraphPad Prism for statistical analysis

**Criteres de succes :**

- ADCC potency of S239D/I332E vs. WT: At 10 µg/mL and E:T 10:1, S239D/I332E shows ≥ 40% specific lysis, and WT shows < 10% specific lysis (p < 0.01, two-way ANOVA with Tukey's HSD).
- FcγRIIIa genotype-dependent effect: At 1 µg/mL and E:T 10:1, NK cells from V/V donors show ≥ 2-fold higher specific lysis with S239D/I332E compared to NK cells from F/F donors (p < 0.05, unpaired t-test).
- FAP expression density: CAFs must express > 10,000 FAP copies/cell as measured by flow cytometry.
- Specificity: Isotype control shows < 5% specific lysis at the highest concentration tested.

- **GO**: All success criteria are met: S239D/I332E shows a robust and significant increase in ADCC over WT, the effect is clearly dependent on FcγRIIIa affinity (V/V > F/F), and the assay is specific.
- **NO-GO**: S239D/I332E fails to show a significant improvement over WT (e.g., <2-fold increase), OR the effect is not genotype-dependent, suggesting the mechanism is not primarily driven by FcγRIIIa affinity.
- **PIVOT**: If the afucosylated variant was predicted to be superior in Phase 1, include it in this phase for comparison. If ADCC is observed but requires very high antibody concentrations (>100 µg/mL), consider exploring a bispecific antibody format in Phase 3.

### 4.3 Phase 3 — Full Experimental Protocol (In Vivo Efficacy & TME Remodeling)

**Objectif**: To rigorously validate the full mechanistic chain in a relevant in vivo model: demonstrating that the Fc-engineered antibody depletes FAP+ CAFs, reduces immunosuppressive cytokines, increases CD8+ T cell infiltration, and inhibits tumor growth, thereby providing publishable proof-of-concept.

**Methodologie**: 1. **In Vivo Model:**
   - Use an immunocompetent mouse model. MC38 (colon carcinoma) or EMT-6 (breast carcinoma) are suitable. Co-inject 5x10^5 tumor cells with 2x10^5 FAP+ CAFs (mouse CAFs, e.g., from a mouse fibroblast line like 3T3, transduced to express mouse FAP) subcutaneously into the flank of C57BL/6 (for MC38) or BALB/c (for EMT-6) mice.
   - Use a surrogate anti-mouse FAP antibody with the S239D/I332E mutation (or the mouse equivalent, mS239D/I332E) to ensure proper Fc-FcγRIIIa interaction. Control groups: (1) Vehicle (PBS), (2) WT anti-mouse FAP antibody, (3) S239D/I332E anti-mouse FAP antibody, (4) Isotype control (S239D/I332E).
2. **Treatment Protocol:**
   - When tumors reach ~100 mm³, randomize mice into groups (n=8-10 per group).
   - Administer antibodies intraperitoneally at 10 mg/kg, twice weekly for 3 weeks.
   - Measure tumor volume every 2-3 days with calipers.
3. **TME Analysis (Endpoint):**
   - At the end of the study (day 21), euthanize mice and harvest tumors.
   - **Flow Cytometry:** Dissociate tumors into single-cell suspensions. Stain for CD45, CD3, CD4, CD8, FoxP3, NK1.1, and FAP (to assess CAF depletion). Analyze on a flow cytometer (e.g., BD LSRFortessa).
   - **Cytokine/Chemokine Analysis:** Use a multiplex bead-based assay (e.g., LEGENDplex, BioLegend) on tumor lysates to quantify CCL2, WNT2, and other key cytokines (e.g., IFN-γ, TNF-α, TGF-β). Normalize to total protein (BCA assay).
   - **Immunohistochemistry (IHC):** Fix and paraffin-embed tumor sections. Stain for CD8, FoxP3, and FAP. Quantify cell densities (cells/mm²) using digital pathology software (e.g., QuPath).
4. **Statistical Analysis:**
   - Tumor growth curves: Two-way ANOVA with repeated measures, followed by Dunnett's post-hoc test.
   - Endpoint analyses (cell densities, cytokine levels): One-way ANOVA with Dunnett's post-hoc test. P < 0.05 considered significant.

- Cout: €40,000-100,000 (mainly for animal housing, antibodies, and multiplex kits)
- Duree: 6-9 months
- Equipement: Animal facility (IVC cages), Caliper for tumor measurement, Flow cytometer (e.g., BD LSRFortessa), Tissue homogenizer, Multiplex plate reader (e.g., Bio-Plex 200), Microtome and slide scanner for IHC
- Logiciels: GraphPad Prism, FlowJo, QuPath (for IHC quantification)

**Criteres de succes :**

- Tumor growth inhibition: ≥ 50% reduction in mean tumor volume (mm³) in the S239D/I332E group compared to the vehicle control group at day 21 (p < 0.01, one-way ANOVA with Dunnett's post-hoc test).
- CAF depletion: ≥ 50% reduction in the percentage of FAP+ cells within the CD45- stromal population in the S239D/I332E group compared to vehicle control.
- CD8+ T cell infiltration: ≥ 2-fold increase in intratumoral CD8+ T cell density (cells/mm²) in the S239D/I332E group compared to vehicle control (p < 0.05).
- Treg density: ≥ 50% reduction in intratumoral Treg density (cells/mm²) in the S239D/I332E group compared to vehicle control.
- Cytokine reduction: ≥ 50% reduction in CCL2 and WNT2 levels (pg/mg total protein) in the S239D/I332E group compared to vehicle control.

- **GO**: All primary success criteria are met: significant tumor growth inhibition, CAF depletion, increased CD8+ T cells, decreased Tregs, and reduced immunosuppressive cytokines.
- **NO-GO**: Tumor growth inhibition is < 30%, OR there is no significant CAF depletion, OR no significant change in CD8+ T cell infiltration, indicating the mechanism is not operative in vivo.
- **PIVOT**: If CAF depletion is observed but no T cell infiltration, the mechanism may be solely CAF killing without immune remodeling. Pivot to investigate direct effects on tumor cells or combine with anti-PD-1 in a follow-up study. If systemic toxicity is observed (e.g., weight loss), pivot to a lower dose or a modified dosing schedule.

### 4.4 Quick start : comment demarrer aujourd'hui

- **Peut demarrer maintenant**: Oui
- **Premiere action**: Download the PDB structures of FAP (1Z68) and FcγRIIIa (3AY4), and initiate a literature search on GEO/TCGA for colorectal and pancreatic cancer datasets with matched RNA-seq and clinical outcome data.
- **Outils**: PyMOL (free for academics), Python with NumPy/SciPy, R with Seurat/CIBERSORTx (free), Access to GEO/TCGA databases
- **Donnees ouvertes**: TCGA (portal.gdc.cancer.gov), GEO (ncbi.nlm.nih.gov/geo), Human Protein Atlas (proteinatlas.org), Human Cell Atlas (humancellatlas.org)

## 5. Analyse d'impact

### 5.1 Impact scientifique

Novelty score: 0.8/1.0 (novel)

### 5.2 Applications industrielles et marche

**Score industriel**: 6.8/10

**Forces :**
- Cible hautement validee cliniquement : FAP est une cible stromale bien etablie avec des anticorps (FAP-IL2v, RO6874281) deja en essais cliniques par Roche, ce qui valide l'accessibilite de la cible et reduit le risque de developpement technique.
- Differentiation claire via l'ingenierie Fc : L'approche S239D/I332E (technologie 'DELTA' de Xencor) est une plateforme eprouvee (ex: mogamulizumab) qui peut etre licenciee, offrant un avantage competitif immediat par rapport aux anti-FAP non optimises pour l'ADCC.
- Marche adressable substantiel dans les tumeurs stroma-riches : Le ciblage du TME immunosuppresseur s'applique aux cancers du sein triple negatif, du pancreas (PDAC) et colorectal, representant un marche mondial de plusieurs milliards de dollars en therapies combinees.

**Faiblesses :**
- Toxicite potentielle sur les FAP+ non tumoraux : FAP est exprime dans la fibrose de cicatrisation, les tissus musculo-squelettiques et la moelle osseuse ; une depletion systemique via ADCC renforcee pourrait induire des toxicites hors-cible (cachexie, fibrose) non capturees par les modeles murins.
- Dependance a la densite de cellules NK intratumorales : L'hypothese repose sur la presence de NK fonctionnels dans le TME, or les tumeurs stroma-riches sont souvent 'cold' avec une infiltration NK faible ; le protocole ne demontre pas de strategie pour recruter/activer les NK in situ.
- Concurrence des approches bispecifiques et CAR-M : Les bispecifiques FAP/CD3 (ex: FAP-4-1BBL) et les CAR-T ciblant FAP sont deja en developpement preclinique avance, offrant une depletion des CAFs potentiellement plus robuste que l'ADCC seule.

**Recommandation**: Poursuivre le developpement mais pivoter vers une strategie de 'best-in-class' en combinant l'ingenierie Fc avec un format bispecifique (ex: anti-FAP x anti-NK engager) pour maximiser la depletion des CAFs et surmonter la faible infiltration NK. Cibler en priorite les indications PDAC et TNBC ou le stroma est dense, et initier des discussions de licensing avec des acteurs comme Roche ou Xencor des la validation in vivo du protocole Phase 3, afin de partager le risque de toxicite et accelerer l'acces aux essais cliniques.

### 5.3 Opportunites de financement

| Programme | Agence | Fit | Budget type | Taux succes |
|-----------|--------|-----|-------------|-------------|
| ANR Jeune Chercheuse - Jeune Chercheur (JCJC) 2026 | Agence Nationale de la Recherche (ANR) - France | 0.85 | €150k - €300k | Environ 15-20% pour les jeunes chercheurs |
| EIC Pathfinder Open 2026 | European Innovation Council (EIC) - Commission Européenne | 0.65 | Jusqu'à €4 millions (mais des projets plus petits sont acceptés) | Environ 5-7% |
| Fondation ARC - Appel à projets 'Immunothérapies innovantes' 2025 | Fondation ARC pour la recherche sur le cancer (France) | 0.75 | €100k - €200k | Environ 20-25% |

- **ANR Jeune Chercheuse - Jeune Chercheur (JCJC) 2026** (Agence Nationale de la Recherche (ANR) - France): Ce programme est idéal pour un chercheur indépendant souhaitant financer une preuve de concept à risque. Le budget demandé (€22k-117k) est compatible avec un financement JCJC, et le caractère mécanistique du projet (validation d'une hypothèse de biologie) correspond parfaitement aux critères d'excellence scientifique de l'ANR. Le dossier peut être porté par un seul PI, ce qui évite la complexité d'un consortium à ce stade précoce.
- **EIC Pathfinder Open 2026** (European Innovation Council (EIC) - Commission Européenne): Ce programme finance des idées radicalement nouvelles à haut risque (TRL 2-4). L'ingénierie d'anticorps Fc pour cibler le stroma tumoral est une approche disruptive qui pourrait intéresser l'EIC. Cependant, le budget demandé est très inférieur au budget type, ce qui nécessiterait de redimensionner le projet (ajouter des partenaires, des études de toxicité, un volet propriété intellectuelle). Le programme exige un fort potentiel d'innovation de rupture et une stratégie de mise sur le marché, ce qui est prématuré à ce stade.
- **Fondation ARC - Appel à projets 'Immunothérapies innovantes' 2025** (Fondation ARC pour la recherche sur le cancer (France)): Cet appel est spécifiquement dédié aux approches thérapeutiques émergentes en immunothérapie. Le ciblage des CAF via l'ADCC est une thématique prioritaire pour la fondation, qui finance des projets à risque mais avec un fort potentiel translationnel. Le budget demandé (€117k) est dans la fourchette haute de l'appel, et le protocole en 3 phases avec des jalons GO/NO-GO est un format apprécié. La fondation valorise les collaborations avec des plateformes de recherche clinique, ce qui pourrait être un point à renforcer dans le consortium.

**Recommandation financement**: Cibler en priorité l'ANR Jeune Chercheuse (JCJC) ou un appel flash de la fondation ARC pour obtenir un financement rapide de la Phase 1 et 2 (preuve de concept in vitro). En parallèle, utiliser les données de la Phase 2 pour construire un consortium avec un laboratoire d'immunologie tumorale et un CHU pour soumettre un projet Horizon Europe (Cluster 1 ou EIC Transition) en année 2, après avoir démontré l'efficacité in vivo.

## 6. Panel Review Summary

| Reviewer | Score | Verdict | Point cle |
|----------|-------|---------|-----------|
| methodologist | 7.8/10 | accept | The protocol demonstrates exceptional rigor through its phased, de-risking approach. The inclusion of a Phase 1 in silico validation with explicit GO/NO-GO criteria (e.g., ≥4-fold predicted ADCC improvement, therapeutic index ≥10:1) is a sophisticated and cost-effective strategy that prioritizes mechanistic feasibility before committing to expensive in vivo studies. |
| domain_expert | 7.8/10 | accept | The hypothesis demonstrates strong theoretical coherence by directly transposing a validated paradigm from immuno-oncology: the Fc-engineering of antibodies to deplete immunosuppressive populations via ADCC, as established with anti-CTLA-4 and anti-4-1BB antibodies, onto the clinically relevant target of FAP+ CAFs. This is a logical and well-supported extension of the state of the art. |
| contrarian | 3.5/10 | weak_reject | The proposal correctly identifies a genuine biological barrier: the low baseline ADCC against stromal cells due to inhibitory signals (e.g., HLA-E) and the physical density of the ECM. Targeting CAFs via Fc-engineering is a logical extension of existing checkpoint and anti-angiogenic strategies, and the S239D/I332E mutation is a well-validated Fc enhancement approach. |
| industrialist | 6.8/10 | accept | Cible hautement validee cliniquement : FAP est une cible stromale bien etablie avec des anticorps (FAP-IL2v, RO6874281) deja en essais cliniques par Roche, ce qui valide l'accessibilite de la cible et reduit le risque de developpement technique. |
| funding_strategist | 6.8/10 | weak_accept | Mécanisme clairement articulé avec une chaîne causale testable (Fc-engineering → ADCC → déplétion CAF → remodelage TME), ce qui facilite la rédaction d'un narratif scientifique convaincant pour les évaluateurs. |

**Consensus score**: 6.5/10
**Verdict final**: publish_brief

### 6.1 Consensus

- L'ingénierie Fc S239D/I332E est une approche validée et pertinente pour augmenter l'affinité FcγRIIIa et potentiellement surmonter la faible susceptibilité à l'ADCC des cellules stromales comme les CAF FAP+.
- La conception expérimentale en trois phases avec des critères GO/NO-GO quantitatifs et prédéfinis est une force méthodologique majeure, permettant un dérisquage efficace et une évaluation rigoureuse de l'hypothèse.
- Le ciblage des CAF FAP+ représente une stratégie de différenciation pertinente dans le microenvironnement tumoral immunosuppresseur, avec un potentiel translationnel et un marché adressable substantiel dans les tumeurs stroma-riches.

### 6.2 Points de desaccord

- Le principal désaccord concerne la pertinence physiologique du modèle et l'hypothèse sous-jacente de la présence et de la fonctionnalité des cellules NK dans les tumeurs desmoplasiques 'cold', que le Contrarian juge comme une faille rédhibitoire, tandis que les autres experts la considèrent comme un risque à atténuer.
- Un second point de divergence porte sur la suffisance de la déplétion des CAF pour induire le remodelage immunitaire prédit, le Contrarian et l'expert domaine soulignant la redondance des sources de cytokines (ex: TAMs) et les mécanismes compensatoires potentiels, alors que le méthodologiste et l'industrialiste voient la chaîne causale comme plausible et testable.
- Enfin, la question de la toxicité potentielle hors-cible (FAP+ non tumoraux) est un sujet de préoccupation partagé, mais avec des implications différentes : l'industrialiste et le funding strategist y voient un risque de développement à gérer, tandis que le Contrarian le considère comme un obstacle potentiellement non résolu par le protocole actuel.

### 6.3 Critical path

Le facteur le plus déterminant pour le succès de cette hypothèse est la validation expérimentale de la présence et de la fonctionnalité des cellules NK dans le microenvironnement tumoral desmodplasique ciblé. Sans une démonstration que les NK peuvent être recrutés et activés in situ (ou que leur absence peut être compensée), l'efficacité de l'ADCC, même optimisée par l'ingénierie Fc, reste une solution à un problème qui pourrait ne pas exister dans les tumeurs 'cold'.

**Recommandation finale**: Le panel recommande la publication du brief pour validation expérimentale, car l'hypothèse est mécanistiquement solide, méthodologiquement rigoureuse et stratégiquement pertinente. Cependant, la publication doit être conditionnée à une révision intégrant explicitement les préoccupations majeures : 1) une justification statistique formelle (power analysis) pour la phase in vivo, 2) un plan détaillé pour évaluer la densité et la fonction des NK dans les modèles tumoraux, et 3) une stratégie pour surveiller et atténuer la toxicité systémique potentielle. Le protocole doit également inclure des contrôles pour la redondance des cytokines et les mécanismes compensatoires. Ces ajouts sont essentiels pour transformer une hypothèse prometteuse en un plan expérimental capable de produire des données concluantes et de répondre aux critiques les plus sévères.

## 7. Gap Manifest residuel

### 7.1 Data gaps

- Direct experimental evidence for the susceptibility of FAP+ CAFs to NK cell-mediated ADCC is still missing. No paper in the provided list directly tests this.
- No data exists on the specific Fc-engineering of an anti-FAP antibody (e.g., afucosylation, S239D/I332E mutations) and its resulting ADCC activity against CAFs.
- The expression levels of NK cell ligands (e.g., MICA/B, CD155) and MHC-I on FAP+ CAFs, which would determine their susceptibility to NK cell killing, are not characterized in the provided papers.

### 7.2 Competence gaps

- Phase 1: Computational biology / bioinformatics
- Phase 1: Structural biology / molecular modeling
- Phase 1: Basic immunology knowledge
- Phase 1: Python/R programming
- Phase 2: Primary cell culture (fibroblasts, NK cells)
- Phase 2: Flow cytometry
- Phase 2: Cytotoxicity assays (LDH or 51Cr release)
- Phase 2: Basic molecular biology (genotyping)
- Phase 2: Statistical analysis
- Phase 3: Animal handling and in vivo efficacy studies
- Phase 3: Tumor dissociation and flow cytometry
- Phase 3: Multiplex immunoassays
- Phase 3: Immunohistochemistry and digital pathology
- Phase 3: Advanced statistical analysis

### 7.3 Epistemic gaps

- We do not know the threshold of FAP expression density required for efficient NK cell-mediated ADCC.
- We do not know whether the in vitro ADCC activity will translate to in vivo efficacy given the hypoxic and metabolically challenging TME.
- We do not know the extent to which CAF depletion will lead to compensatory mechanisms from other fibroblast subsets.

## References

[1] D. Ha, A. Tanaka, et al. (2018). *Differential control of human Treg and effector T cells in tumor immunity by Fc-engineered anti–CTLA-4 antibody*. DOI: [10.1073/pnas.1812186116](https://doi.org/10.1073/pnas.1812186116)
[2] S. Buchan, L. Dou, et al. (2018). *Antibodies to Costimulatory Receptor 4‐1BB Enhance Anti‐tumor Immunity via T Regulatory Cell Depletion and Promotion of CD8 T Cell Effector Function*. DOI: [10.1016/j.immuni.2018.09.014](https://doi.org/10.1016/j.immuni.2018.09.014)
[3] Shichuan Hu, Jinhu Ma, et al. (2021). *Engineered exosome-like nanovesicles suppress tumor growth by reprogramming tumor microenvironment and promoting tumor ferroptosis.*. DOI: [10.1016/j.actbio.2021.09.003](https://doi.org/10.1016/j.actbio.2021.09.003)
[4] Rongyuan Wei, Junquan Song, et al. (2025). *FAP upregulates PD-L1 expression in cancer-associated fibroblasts to exacerbate T cells dysfunction and suppress anti-tumor immunity.*. DOI: [10.1016/j.canlet.2025.217475](https://doi.org/10.1016/j.canlet.2025.217475)
[5] A. Kraxner, F. Braun, et al. (2024). *Investigating the complex interplay between fibroblast activation protein α-positive cancer associated fibroblasts and the tumor microenvironment in the context of cancer immunotherapy*. DOI: [10.3389/fimmu.2024.1352632](https://doi.org/10.3389/fimmu.2024.1352632)
[6] Hsing Hwa Lee, Z. Al-Ogaili (2025). *Fibroblast activation protein and the tumour microenvironment: challenges and therapeutic opportunities*. DOI: [10.3389/or.2025.1617487](https://doi.org/10.3389/or.2025.1617487)
[7] Vidhi Khanna, Hyunjoon Kim, et al. (2021). *Novel TLR 7/8 agonists for improving NK cell mediated antibody-dependent cellular cytotoxicity (ADCC)*. DOI: [10.1038/s41598-021-83005-6](https://doi.org/10.1038/s41598-021-83005-6)
[8] H. Vietzen, Vera Danklmaier, et al. (2022). *High-affinity FcγRIIIa genetic variants and potent NK cell-mediated antibody-dependent cellular cytotoxicity (ADCC) responses contributing to severe COVID-19*. DOI: [10.1016/j.gim.2022.04.005](https://doi.org/10.1016/j.gim.2022.04.005)
[9] Femke A I Ehlers, Nicky A. Beelen, et al. (2021). *ADCC-Inducing Antibody Trastuzumab and Selection of KIR-HLA Ligand Mismatched Donors Enhance the NK Cell Anti-Breast Cancer Response*. DOI: [10.3390/cancers13133232](https://doi.org/10.3390/cancers13133232)
[10] Ramesh Mukkamala, Daniel J. Carlson, et al. (2024). *Design of a Fibroblast Activation Protein–Targeted Radiopharmaceutical Therapy with High Tumor–to–Healthy-Tissue Ratios*. DOI: [10.2967/jnumed.124.267756](https://doi.org/10.2967/jnumed.124.267756)
[11] Elizabeth K Duperret, A. Trautz, et al. (2017). *Alteration of the tumor stroma using a consensus DNA vaccine targeting Fibroblast Activation Protein (FAP) synergizes with anti-tumor vaccine therapy in mice*. DOI: [10.1158/1078-0432.CCR-17-2033](https://doi.org/10.1158/1078-0432.CCR-17-2033)
[12] Joshua D Freedman, M. Duffy, et al. (2018). *An Oncolytic Virus Expressing a T-cell Engager Simultaneously Targets Cancer and Immunosuppressive Stromal Cells.*. DOI: [10.1158/0008-5472.CAN-18-1750](https://doi.org/10.1158/0008-5472.CAN-18-1750)
[13] Marc Wehrli, Samantha Guinn, et al. (2024). *Mesothelin CAR T-cells secreting anti-FAP/anti-CD3 molecules efficiently target pancreatic adenocarcinoma and its stroma*. DOI: [10.1158/1078-0432.CCR-23-3841](https://doi.org/10.1158/1078-0432.CCR-23-3841)
[14] Reyisa Bughda, P. Dimou, et al. (2021). *Fibroblast Activation Protein (FAP)-Targeted CAR-T Cells: Launching an Attack on Tumor Stroma*. DOI: [10.2147/ITT.S291767](https://doi.org/10.2147/ITT.S291767)

## Annexes

### A. Detailed Reviewer Reports

#### Methodologist

- **Score**: 7.8/10 | **Verdict**: accept | **Confidence**: 0.9
- **Strengths**: The protocol demonstrates exceptional rigor through its phased, de-risking approach. The inclusion of a Phase 1 in silico validation with explicit GO/NO-GO criteria (e.g., ≥4-fold predicted ADCC improvement, therapeutic index ≥10:1) is a sophisticated and cost-effective strategy that prioritizes mechanistic feasibility before committing to expensive in vivo studies.; The experimental design is methodologically sound, incorporating multiple critical controls (isotype, no-antibody, positive control for NK function) and addressing a key biological variable—FcγRIIIa (CD16a) V/F polymorphism—by genotyping donors and stratifying them into V/V and F/F groups. This directly tests the mechanistic link between Fc affinity and ADCC, strengthening causal inference.; The hypothesis is exceptionally well-formulated and falsifiable, with clear, quantitative, and pre-specified success bounds for each prediction (e.g., ≥40% specific lysis for S239D/I332E vs. <10% for WT). This commitment to pre-registering effect sizes and statistical thresholds (p-values, specific tests) is a hallmark of high-quality, reproducible science.
- **Weaknesses**: The protocol lacks a pre-specified statistical power analysis for the in vivo Phase 3 study. While n=8-10 per group is a common convention, it is not justified. A formal power calculation based on expected effect sizes and variance from pilot data would significantly strengthen the design and ensure the study is adequately powered to detect the hypothesized ≥50% reduction in tumor volume.; The in vivo model (MC38/EMT-6 co-injected with FAP+ CAFs) is a significant simplification of the complex human TME. It does not fully recapitulate the desmoplastic stroma or the chronic, suppressive microenvironment found in human tumors. This limits the generalizability of the findings and the conclusions that can be drawn about TME remodeling.; The protocol does not adequately address the potential for off-target effects or systemic toxicity of the Fc-engineered antibody in vivo. While the Phase 1 in silico assessment is a good first step, the Phase 3 protocol lacks explicit monitoring for adverse events (e.g., body weight, organ histology) or a pre-defined humane endpoint for toxicity, which is a critical ethical and methodological omission.
- **Questions**: What is the a priori statistical power calculation for the Phase 3 in vivo study? Please provide the expected effect size, standard deviation, and alpha level used to justify the n=8-10 per group, or state whether the study is powered to detect a smaller, clinically relevant effect if the hypothesized ≥50% reduction is overly optimistic.; How will you ensure the validity and reliability of the IHC quantification for CD8+ and FoxP3+ cells? Will you use automated, unbiased digital pathology algorithms, and will the pathologist be blinded to the treatment groups to mitigate measurement bias?; Given the known variability in primary NK cell activity, what specific measures are in place to control for inter-donor and intra-donor variability in the ADCC assays (e.g., using a single batch of IL-2-activated NK cells, normalizing to a reference standard, or using a larger panel of donors for a more robust mixed-effects model)?
- **Recommendation**: This is a well-designed, logically structured protocol that demonstrates a strong understanding of experimental methodology. The phased approach with clear go/no-go criteria is exemplary. I recommend acceptance, contingent on the addition of a formal power analysis for the in vivo phase and explicit monitoring for systemic toxicity. Addressing the limitations of the in vivo model in the discussion will also be crucial for a balanced interpretation of the results.

#### Domain Expert

- **Score**: 7.8/10 | **Verdict**: accept | **Confidence**: 0.85
- **Strengths**: The hypothesis demonstrates strong theoretical coherence by directly transposing a validated paradigm from immuno-oncology: the Fc-engineering of antibodies to deplete immunosuppressive populations via ADCC, as established with anti-CTLA-4 and anti-4-1BB antibodies, onto the clinically relevant target of FAP+ CAFs. This is a logical and well-supported extension of the state of the art.; The proposed causal chain is mechanistically plausible and well-articulated. The use of the S239D/I332E double mutation is a standard, high-impact engineering approach to enhance FcγRIIIa affinity, and the downstream consequences (CAF depletion → reduced WNT2/CCL2 → increased CD8/Treg ratio) are supported by a credible, albeit partially indirect, body of literature.; The hypothesis correctly identifies and addresses a critical bottleneck in the field: the low intrinsic ADCC susceptibility of stromal cells. By framing the problem as an Fc-FcγRIIIa affinity issue, it proposes a specific, testable solution that goes beyond simply targeting FAP and leverages a deep understanding of NK cell biology and antibody engineering.
- **Weaknesses**: The hypothesis may underestimate the physical and biological barriers within the tumor microenvironment. The assumption that NK cells are present and fully functional is a significant oversimplification, as tumors often exhibit NK cell exclusion, dysfunction, and metabolic suppression. The hypoxic and acidic TME could severely blunt the efficacy of the proposed ADCC mechanism, a key 'known unknown' that is not adequately addressed in the formulation.; The claim that CAF depletion will lead to a >50% reduction in cytokines and a subsequent decrease in Treg recruitment and T cell exhaustion is an oversimplification of a highly redundant system. The hypothesis does not sufficiently account for potential compensatory mechanisms from other stromal subsets (e.g., FAP-negative CAFs, smooth muscle actin-positive myofibroblasts) or the direct physical barrier function of the extracellular matrix, which may not be resolved by simple cell depletion.; The potential for on-target/off-tumor toxicity is a major concern that is acknowledged but not deeply integrated into the mechanistic model. FAP is expressed on bone marrow stromal cells and in some normal tissues during wound healing. The hypothesis lacks a discussion of how the enhanced ADCC activity might exacerbate this toxicity, which is a critical hurdle for clinical translation and a key point of differentiation from T-cell engagers.
- **Questions**: Given the known metabolic and hypoxic constraints of the TME, what specific strategies (e.g., combination with TGF-β blockade, metabolic modulators, or NK cell engagers) could be employed to ensure that intratumoral NK cells retain the metabolic fitness and cytotoxic capacity required to execute ADCC effectively?; What is the precise threshold of FAP surface density on CAFs required to trigger efficient ADCC, and how does this compare across different tumor types and stages? Does the S239D/I332E mutation lower this threshold sufficiently to overcome the inherent resistance of stromal cells, and can this be validated using quantitative immunofluorescence and single-cell proteomics?; How would the proposed mechanism account for the potential of CAF depletion to induce a reactive wound-healing response, leading to the expansion of other immunosuppressive fibroblast subsets or increased matrix deposition, and what biomarkers would be used to monitor for such compensatory resistance?
- **Recommendation**: This is a well-formulated and novel hypothesis that merits further investigation. The proposed mechanism is a logical and exciting extension of current Fc-engineering strategies to a challenging and important target cell population. However, the proposal should be refined to explicitly address the significant biological barriers posed by the TME and the potential for compensatory resistance. I recommend focusing initial validation on in vitro 3D co-culture systems that better mimic the physical and metabolic constraints of the tumor stroma, and to include a detailed plan for monitoring systemic toxicity in future in vivo studies.

#### Contrarian

- **Score**: 3.5/10 | **Verdict**: weak_reject | **Confidence**: 0.85
- **Strengths**: The proposal correctly identifies a genuine biological barrier: the low baseline ADCC against stromal cells due to inhibitory signals (e.g., HLA-E) and the physical density of the ECM. Targeting CAFs via Fc-engineering is a logical extension of existing checkpoint and anti-angiogenic strategies, and the S239D/I332E mutation is a well-validated Fc enhancement approach.; The authors explicitly acknowledge a key unknown: the threshold of FAP density required for NK-mediated ADCC. This is a testable and critical parameter that, if addressed, would significantly de-risk the hypothesis.
- **Weaknesses**: FAIL REASON #1: The 'NK cell recruitment' assumption is a fatal, untested leap. The hypothesis states NK cells must be present or recruited, but provides no mechanism. CAF-rich desmoplastic tumors (e.g., PDAC, breast cancer) are notoriously 'cold' with very low NK cell infiltration. The in vitro ADCC assay (Prediction 1) uses an E:T ratio of 10:1, which is physiologically irrelevant for a tumor where NK cells represent <1% of the immune infiltrate. Even if ADCC works in a dish, the effector cells are simply absent in vivo.; FAIL REASON #2: The causal chain is built on a 'one-way street' fallacy. The hypothesis assumes CAF depletion reduces WNT2 and CCL2, leading to CD8+ T cell infiltration. However, CAFs are not the sole source of these cytokines. Tumor-associated macrophages (TAMs) and myeloid-derived suppressor cells (MDSCs) also secrete CCL2. Depleting CAFs will likely lead to compensatory upregulation of CCL2 from these other populations, blunting the predicted >50% reduction. The model ignores stromal redundancy and the dynamic, reciprocal signaling between CAFs and immune cells.; FAIL REASON #3: The proposed in vivo model (Prediction 3) is a confounded artifact. Co-injecting FAP+ CAFs with MC38 or HT-29 cells into a subcutaneous xenograft creates an artificial, NK-cell-poor environment in an immunodeficient mouse (if using nude or NSG). You cannot measure CD8+ T cell infiltration in a mouse lacking T cells. If using a syngeneic model (e.g., MC38 in C57BL/6), the mouse FcγRIIIa has different affinity for the human Fc domain, and the S239D/I332E mutation may not translate across species. The model is either immunologically irrelevant or mechanistically invalid.
- **Questions**: What is the intratumoral density of NK cells in the specific tumor type you target, and can you demonstrate that the Fc-engineered antibody promotes NK cell recruitment from the periphery, not just activation of pre-existing cells?; If CAF depletion reduces CCL2 by only 20% (not the predicted 50%) due to compensation by TAMs, does your downstream prediction of Treg reduction and CD8+ infiltration still hold? Provide a sensitivity analysis of your cytokine threshold.; How do you rule out on-target/off-tumor toxicity to FAP+ bone marrow stromal cells, which are known to be essential for hematopoiesis? Your hypothesis states this as a 'key assumption' but provides no mitigation strategy, and a 10 mg/kg dose is likely to cause systemic toxicity.
- **Recommendation**: The authors must first demonstrate NK cell presence and function in a relevant desmoplastic model (e.g., KPC mouse or human PDAC explants). They should use a syngeneic model with a mouse-reactive anti-FAP antibody (or a Fc-engineered mouse IgG) to test the causal chain. Furthermore, they must perform a 'cell depletion' experiment using a FAP-Cre/DTR mouse model to prove that CAF depletion alone (without Fc-engineering) is sufficient to achieve the predicted cytokine reduction and T cell infiltration. Without this, the Fc-engineering is a solution looking for a problem that may not exist in vivo.

#### Industrialist

- **Score**: 6.8/10 | **Verdict**: accept | **Confidence**: 0.72
- **Strengths**: Cible hautement validee cliniquement : FAP est une cible stromale bien etablie avec des anticorps (FAP-IL2v, RO6874281) deja en essais cliniques par Roche, ce qui valide l'accessibilite de la cible et reduit le risque de developpement technique.; Differentiation claire via l'ingenierie Fc : L'approche S239D/I332E (technologie 'DELTA' de Xencor) est une plateforme eprouvee (ex: mogamulizumab) qui peut etre licenciee, offrant un avantage competitif immediat par rapport aux anti-FAP non optimises pour l'ADCC.; Marche adressable substantiel dans les tumeurs stroma-riches : Le ciblage du TME immunosuppresseur s'applique aux cancers du sein triple negatif, du pancreas (PDAC) et colorectal, representant un marche mondial de plusieurs milliards de dollars en therapies combinees.
- **Weaknesses**: Toxicite potentielle sur les FAP+ non tumoraux : FAP est exprime dans la fibrose de cicatrisation, les tissus musculo-squelettiques et la moelle osseuse ; une depletion systemique via ADCC renforcee pourrait induire des toxicites hors-cible (cachexie, fibrose) non capturees par les modeles murins.; Dependance a la densite de cellules NK intratumorales : L'hypothese repose sur la presence de NK fonctionnels dans le TME, or les tumeurs stroma-riches sont souvent 'cold' avec une infiltration NK faible ; le protocole ne demontre pas de strategie pour recruter/activer les NK in situ.; Concurrence des approches bispecifiques et CAR-M : Les bispecifiques FAP/CD3 (ex: FAP-4-1BBL) et les CAR-T ciblant FAP sont deja en developpement preclinique avance, offrant une depletion des CAFs potentiellement plus robuste que l'ADCC seule.
- **Questions**: Quel est le modele de revenus reel : licencier la technologie Fc (Xencor) et vendre un actif preclinique, ou developper un produit de bout en bout avec un cout de R&D estime a 50-100M€ avant Phase I ?; Comment comptez-vous gerer le risque de toxicite sur les FAP+ non tumoraux dans les etudes reglementaires de toxicologie (NHP) avant de passer en Phase I, sachant que les modeles murins classiques ne sont pas predictifs ?; Quel est le plan de combinaison avec les checkpoint inhibitors (anti-PD-1) pour justifier une valeur ajoutee incrementale, et comment allez-vous convaincre un partenaire pharma que votre ADCC est superieure aux approches de depletion de CAFs deja en essais (ex: FAP-IL2v) ?
- **Recommendation**: Poursuivre le developpement mais pivoter vers une strategie de 'best-in-class' en combinant l'ingenierie Fc avec un format bispecifique (ex: anti-FAP x anti-NK engager) pour maximiser la depletion des CAFs et surmonter la faible infiltration NK. Cibler en priorite les indications PDAC et TNBC ou le stroma est dense, et initier des discussions de licensing avec des acteurs comme Roche ou Xencor des la validation in vivo du protocole Phase 3, afin de partager le risque de toxicite et accelerer l'acces aux essais cliniques.

#### Funding Strategist

- **Score**: 6.8/10 | **Verdict**: weak_accept | **Confidence**: 0.78
- **Strengths**: Mécanisme clairement articulé avec une chaîne causale testable (Fc-engineering → ADCC → déplétion CAF → remodelage TME), ce qui facilite la rédaction d'un narratif scientifique convaincant pour les évaluateurs.; Approche de dérisking pragmatique avec des jalons GO/NO-GO quantitatifs et un budget initial très contenu (€22k-117k), permettant de financer une preuve de concept rapide avant un investissement majeur.; Ciblage des CAF (FAP) plutôt que des cellules tumorales directes, ce qui représente une différenciation stratégique par rapport aux immunothérapies conventionnelles et un angle d'attaque original pour les appels à projets sur le microenvironnement tumoral.
- **Weaknesses**: TRL très faible (TRL 1-2) : l'hypothèse repose entièrement sur des prédictions in silico et des données de littérature non validées par des expériences précliniques, ce qui est un risque majeur pour les programmes de financement axés sur l'impact translationnel.; Absence de partenaires académiques ou cliniques identifiés dans le protocole, ce qui affaiblit la crédibilité pour les appels collaboratifs (Horizon Europe) et limite l'accès aux modèles animaux pertinents et aux échantillons patients pour la stratification.; Le budget demandé (même au maximum de €117k) est en dessous du seuil de visibilité pour la plupart des programmes ANR/ERC, et le calendrier de 10-16 mois est trop court pour générer les données de survie et de toxicité long terme exigées par les évaluateurs.
- **Questions**: Quel est le modèle tumoral in vivo le plus approprié pour valider la déplétion des CAF et le remodelage du TME (ex: modèle syngénique murin avec CAF FAP+ humaines, ou modèle PDX humanisé avec NK cells) ?; Comment allez-vous démontrer que l'effet antitumoral est spécifiquement dû à l'ADCC médiée par les NK et non à d'autres mécanismes (ex: signalisation Fc indépendante, activation des macrophages) ?; Quelle est la stratégie de contrôle pour l'expression de FAP sur les cellules non-CAF (ex: adipocytes, cellules souches mésenchymateuses) afin de justifier un index thérapeutique acceptable ?
- **Recommendation**: Cibler en priorité l'ANR Jeune Chercheuse (JCJC) ou un appel flash de la fondation ARC pour obtenir un financement rapide de la Phase 1 et 2 (preuve de concept in vitro). En parallèle, utiliser les données de la Phase 2 pour construire un consortium avec un laboratoire d'immunologie tumorale et un CHU pour soumettre un projet Horizon Europe (Cluster 1 ou EIC Transition) en année 2, après avoir démontré l'efficacité in vivo.

### B. Semantic Scholar Search Queries

- [novelty] `FAP CARF NK cell ADCC antibody` — Direct search for the core hypothesis combining FAP targeting, NK-mediated killing, and antibody engineering
- [novelty] `anti-FAP antibody afucosylated tumor stroma` — Searches for existing engineered anti-FAP antibodies with enhanced Fc effector function
- [novelty] `fibroblast activation protein Fc engineering immunotherapy` — Broader novelty check on combining FAP targeting with Fc optimization for cancer therapy
- [novelty] `FAP CAF depletion ADCC macrophage` — Checks whether FAP+ CAF depletion via Fc-mediated mechanisms has been previously proposed
- [evidence] `FcgammaR polymorphism antibody efficacy polymorphism` — Establishes the mechanistic link between FcγR variants and antibody-dependent cellular cytotoxicity outcomes
- [evidence] `afucosylated antibody NK cell degranulation` — Provides direct evidence for enhanced ADCC via Fc glycoengineering
- [evidence] `FAP cancer associated fibroblast immunosuppression` — Documents the immunosuppressive role of FAP+ CAFs in the tumor microenvironment
- [evidence] `S239D I332E Fc mutation ADCC` — Validates specific Fc point mutations that enhance FcγRIIIa binding and NK activation
- [cross_domain] `bispecific antibody FAP CD3 tumor stroma` — Explores existing bispecific formats targeting FAP for T-cell engagement, a parallel strategy to NK-cell engagement
- [cross_domain] `CAR macrophage fibroblast activation protein solid tumor` — Examines CAR-based approaches targeting FAP in myeloid cells, providing precedent for innate immune cell redirection
- [cross_domain] `antibody dependent cellular phagocytosis tumor stroma remodeling` — Investigates ADCP as an alternative effector mechanism for stromal cell depletion, informing transferability
- [cross_domain] `Fc engineered antibody Treg depletion tumor` — Looks at Fc engineering applied to deplete immunosuppressive cells (Tregs), a direct precedent for CAF depletion

---

*Generated by SPORE (Systeme de Production d'Opportunites de Recherche par Exploration) on 2026-08-11*