TRKB-based signature identifies high-risk squamous cell carcinoma cases and TRKB blockade reprograms tumor and stromal cells toward suppressive phenotypes.

backgroundCutaneous squamous cell carcinoma (cSCC) is a common age-related cancer, with a subset prone to recurrence and metastasis. Currently, no useful diagnostic biomarkers for high-risk cSCC are available. Based on our previous findings, indicating that age-related changes in the neurotrophin receptor tyrosine kinase-2 (TrkB) axis may promote skin tumorigenesis, this study aims to identify novel cSCC biomarkers and therapeutic targets.

methodsA retrospective analysis was conducted on specimens from patients with in situ or invasive cSCCs using immunohistochemistry to assess the expression of TrkB and specific downstream proteins (i.e., E-cadherin, Yap1, and Notch1). Statistical and machine learning analyses were applied to identify biomarkers that distinguish cSCC subtypes and patient risk groups. In vitro studies involved treating SCC cells, cancer-associated fibroblasts (CAFs), and three-dimensional (3D) SCC models with the TrkB inhibitor ANA-12. Gene and protein expression were analyzed via RTqPCR, immunoblotting, and immunoassays. Functional assays evaluated cell proliferation, migration, and invasion. Secretomes were profiled using cytokine arrays.

resultsProtein expression levels mainly correlated with cSCC types. Our findings indicated that the 'TrkB, E-cadherin, Yap1, Notch1' signature can be a relevant biomarker for both cSCC subtype classification and identification of high-risk cases. Despite the limited sample size, machine learning models demonstrated promising accuracy in differentiating between cSCC classes. Our analysis also highlighted the added value of including stromal markers for classifying high-risk patients. Furthermore, TrkB blockade suppressed tumorigenic traits in TP53-mutant SCC cells, including proliferation, EMT, migration, invasiveness, and disruption of the IL-6/STAT3 signaling loop, while promoting differentiation and senescence through modulation of key players such as p63, Yap1, Notch1, and p21. Data are consistent with a tumor-suppressive effect, thereby promoting tissue homeostasis, especially in physiologically relevant 3D models. Inhibiting TrkB reprogrammed primary CAFs into a less proliferative, migratory, inflammatory, and fibrotic phenotype by simultaneously suppressing key activating pathways, such as β-catenin, Yap1, and Notch1. This aligns with a reduction in their tumor-supportive functions.

conclusionsOur findings provide a basis for improved high-risk patient stratification by highlighting a TrkB-based signature and generating prototype predictive models. Furthermore, they offer promising therapeutic avenues for developing combined targeted interventions to overcome resistance in high-risk patients.