CPHNet: a novel pipeline for anti-HAPE drug screening via deep learning-based Cell Painting scoring

Area of Science:

• Cellular morphology
• High-altitude medicine
• Drug discovery

Background:

• High altitude pulmonary edema (HAPE) is a critical condition caused by rapid ascent to high altitudes.
• Hypoxia triggers cellular changes in the alveolar-capillary barrier, contributing to HAPE pathogenesis.
• Current drug discovery lacks strategies targeting these specific morphological changes.

Purpose of the Study:

• To develop an AI-driven pipeline for identifying anti-HAPE therapeutic agents.
• To focus on morphological alterations in Cell Painting images for drug screening.
• To address the need for effective HAPE interventions based on cellular responses to hypoxia.

Main Methods:

• Generated over 100,000 Cell Painting images of lung cells under hypoxic conditions.
• Utilized a novel segmentation network (SegNet) for subcellular structure detection.
• Developed a hypoxia scoring network (HypoNet) to identify cellular hypoxia status.

Main Results:

• Proposed a deep neural network pipeline (CPHNet) for drug screening.
• Identified ferulic acid (FA) and resveratrol (RES) as promising anti-HAPE compounds.
• Validated the efficacy of FA and RES in ex vivo and in vivo models.

Conclusions:

• Introduced an effective AI and Cell Painting pipeline for screening anti-HAPE agents.
• Provided a novel approach for AI-driven phenotypic drug discovery.
• Opened new avenues for developing targeted HAPE treatments.