Artificial intelligence-driven discovery of YH395A: A novel TGFβR1 inhibitor with potent anti-tumor activity against triple-negative breast cancer.

Characterized by high malignancy and limited treatment efficacy, triple-negative breast cancer (TNBC) remains a clinically challenging subtype within breast cancer classifications, marked by rapid progression and high mortality. Abnormal activation of the transforming growth factor-β (TGFβ) pathway signaling, a pathway integral to tumor progression, metastasis, angiogenesis and immune evasion, is a common feature in a broad spectrum of malignancies. Owing to the restricted effectiveness of first-line interventions including surgical resection, cytotoxic agents, and radiation therapy for TNBC, novel agents that modulate TGFβ activity represent a compelling therapeutic avenue. Herein, we reported the identification and preclinical evaluation of YH395A, a novel tetrahydro-β-carboline derivative derived from the lead compound YR-290 with virtual screening from pseudo molecular library generated by generative deep learning method. In vitro studies demonstrated that YH395A dose-dependently inhibited TNBC cell migration and invasion. In vivo, administration of YH395A not only curtailed metastatic dissemination and prevented the extravasation of breast cancer cells into lung parenchyma in mouse models but also significantly reduced tumor growth in a patient-derived xenograft (PDX) model. Mechanistic analyses indicated that these antitumor effects are mediated via potent inhibition of TGFβ signaling. These cumulative results demonstrate YH395A's viability as a novel therapeutic agent for TNBC, while emphasizing the necessity for expanded preclinical validation studies.