The CdGAP C-terminal pro-invasive domain drives breast cancer metastasis and activates WNT signaling

Technical Summary: The CdGAP C-terminal Pro-invasive Domain Drives Breast Cancer Metastasis and Activates WNT Signaling

Problem Statement
Triple-negative breast cancer (TNBC) and HER2-positive breast cancer are aggressive subtypes characterized by high metastatic potential and limited therapeutic options. While Cdc42 GTPase-activating protein (CdGAP/ARHGAP31) is known to promote metastasis in these contexts, the specific molecular determinants of its pro-metastatic activity remain unclear. Previous studies established CdGAP as an oncogene in HER2+ breast cancer and prostate cancer, yet the functional contribution of its C-terminal region—a domain containing protein-protein interaction sites and post-translational modification motifs but lacking GAP enzymatic activity—has not been defined in the context of breast cancer progression. Furthermore, the mechanism by which CdGAP influences the WNT signaling pathway, a critical driver of tumor stemness and metastasis, remains unexplored.

Methodology
The study employed a multi-faceted approach combining in vivo mouse models, CRISPR/Cas9-mediated genetic engineering in human cell lines, and clinical bioinformatics:

1. Genetic Mouse Models: The authors generated a CdGAP knock-in (ki) mouse model (P668T P670X) mimicking the human Adams-Oliver Syndrome (AOS) truncating mutation, which lacks the C-terminal domain. These mice were crossed with MMTV-Neu transgenic mice to assess tumor initiation, progression, and metastasis in an HER2+ background.
2. Cell Line Engineering: In TNBC cell lines (MDA-MB-231 and HS578T), the authors generated:
   • CdGAP Knock-Out (KO) lines via CRISPR/Cas9.
   • CdGAP Knock-In (KI) lines expressing a C-terminally truncated protein (mimicking the AOS mutation).
   • LIMD1 KO and double KO (LIMD1/CdGAP) lines.
3. Functional Assays: The study utilized a comprehensive suite of assays including:
   • In vivo: Intravascular tumor cell quantification (CTCs), tail vein injection for lung metastasis, and whole-mount mammary gland analysis.
   • In vitro: 3D Matrigel spheroid assays, wound healing, transwell migration/invasion, cell proliferation, and immunofluorescence for focal adhesion dynamics.
   • Molecular: Co-immunoprecipitation to map protein interactions (specifically with LIMD1), Rac1 activation assays, LRP6 surface staining, and Dual Luciferase Reporter assays for WNT/TCF activity.
4. Transcriptomics and Clinical Correlation: RNA-sequencing was performed on KO and KI cells to identify shared transcriptional programs. A CdGAP-dependent WNT gene signature was derived and validated against three independent clinical TNBC cohorts (MBC, Jézéquel, and Brueffer) to assess correlations with metastasis-free and overall survival.

Key Contributions and Results

• The C-terminus is Essential for Metastasis but Not Primary Tumorigenesis: In the MMTV-Neu HER2+ mouse model, deletion of the CdGAP C-terminus did not significantly alter primary tumor formation, histology, or proliferation markers. However, it significantly reduced lung metastasis frequency and circulating tumor cell (CTC) burden, indicating the C-terminus is specifically required for the invasive and metastatic steps of the disease.
• Regulation of TNBC Cell Dynamics: In TNBC cells, loss of the CdGAP C-terminus (KI) or total loss (KO) resulted in reduced proliferation, altered cell morphology (increased cell area, reduced elongation), and enhanced cell-cell adhesion. Consequently, migration, invasion, and lung colonization capabilities were severely impaired.
• Identification of LIMD1 as a C-terminal Interactor: The study identified LIMD1, a LIM-domain protein, as a direct binding partner of the CdGAP C-terminus. This interaction occurs at focal adhesions and requires the arginine residue R1172 on CdGAP. Co-depletion of CdGAP and LIMD1 resulted in additive defects in focal adhesion dynamics and cell migration, suggesting they function cooperatively to regulate cytoskeletal rearrangements.
• Activation of Canonical WNT Signaling: Transcriptomic profiling revealed that CdGAP and its C-terminal domain control a shared gene expression program enriched for WNT signaling. Mechanistically, the CdGAP C-terminus suppresses the expression of KREMEN1, a known inhibitor of WNT signaling. Loss of the C-terminus leads to upregulated KREMEN1, which promotes the endocytosis of the WNT co-receptor LRP6, thereby reducing cell-surface LRP6 availability and dampening canonical WNT/β-catenin signaling.
• Modulation by LIMD1 Binding: Interestingly, the interaction with LIMD1 appears to restrain CdGAP's ability to activate WNT signaling. A CdGAP mutant unable to bind LIMD1 (R1172A) acted as a dominant-positive, increasing WNT transcriptional activity compared to wild-type CdGAP.
• Clinical Prognostic Value: A 18-gene WNT signature dependent on CdGAP was developed. High scores for this signature were significantly associated with shorter time to metastasis and reduced overall survival in multiple independent TNBC patient cohorts.

Significance
The paper establishes the CdGAP C-terminal region as a distinct "pro-invasive domain" (PID) that drives breast cancer metastasis through a GAP-independent mechanism. The study uncovers a novel axis where CdGAP cooperates with LIMD1 at focal adhesions to regulate cell morphology and migration, while simultaneously suppressing the WNT inhibitor KREMEN1 to sustain canonical WNT signaling.

The authors conclude that the CdGAP C-terminus is a critical therapeutic target in aggressive breast cancer subtypes (TNBC and HER2+). By linking focal adhesion dynamics to WNT pathway activation, these findings provide a mechanistic rationale for targeting the CdGAP-LIMD1 interaction or the downstream WNT pathway to inhibit metastasis and improve patient survival in TNBC. The study highlights that while CdGAP acts as an oncogene, its C-terminal truncation (as seen in AOS) results in a loss-of-function phenotype regarding metastasis, contrasting with the gain-of-function often seen in complete CdGAP overexpression.