High-Content Image-Based Screening and Deep Learning for the Detection of Anti-Inflammatory Drug Leads.

Area of Science:

• Immunology
• Pharmacology
• Computational Biology

Background:

• The innate immune response, particularly macrophage activation by lipopolysaccharide (LPS), is a key target for anti-inflammatory drug discovery.
• Identifying novel compounds that modulate this response requires efficient screening methods.

Purpose of the Study:

• To develop and validate a high-content, image-based screening platform for identifying novel anti-inflammatory compounds.
• To leverage deep learning for analyzing cellular morphology changes induced by LPS stimulation.

Main Methods:

• A high-content image-based screen using murine macrophages (RAW264.7) stimulated with LPS.
• Development of a deep neural network classifier trained on fluorescence microscopy images to detect LPS-induced phenotypes.
• Validation of top hits using linear classifiers and a multiplexed cytokine secretion assay.

Main Results:

• Screened 2,259 bioactive compounds, identifying 12 validated hits that modulated cytokine expression.
• Seven hits were identified as allosteric inhibitors of mitogen-activated protein kinase kinase (MEK1/2).
• The deep learning model effectively distinguished between unstimulated and LPS-stimulated macrophages.

Conclusions:

• The developed deep learning morphological assay is effective in identifying compounds that modulate the innate immune response to LPS.
• This platform shows promise for accelerating the discovery of new anti-inflammatory drug leads.