Integrating inflammatory biomarker analysis and artificial-intelligence-enabled image-based profiling to identify

Shan Yu1, Alexandr A Kalinin2, Maria D Paraskevopoulou3

  • 1Takeda Development Center Americas, Inc., San Diego, CA 92121, USA.

Cell Chemical Biology
|July 12, 2023
PubMed

Insights

Researchers identified potential treatments for intestinal fibrosis using Cell Painting and AI. This novel approach screens compounds to reverse fibrotic cell phenotypes, offering new hope for treating this debilitating condition.

Area of Science:

  • Gastroenterology
  • Drug Discovery
  • Computational Biology

Background:

  • Intestinal fibrosis, a complication of inflammatory bowel disease, causes significant morbidity like stenosis and obstruction.
  • Current treatment options for intestinal fibrosis are limited, hindering effective patient management.
  • The absence of screenable cellular phenotypes impedes drug discovery efforts for intestinal fibrosis.

Purpose of the Study:

  • To develop and implement a novel phenotypic screening platform for identifying therapeutic compounds for intestinal fibrosis.
  • To leverage Cell Painting and machine learning to identify small molecules capable of reversing intestinal myofibroblast activation.
  • To discover new drug targets and classes for treating intestinal fibrosis through high-throughput screening.

Main Methods:

  • Utilized Cell Painting, a scalable image-based morphology assay, to analyze cellular phenotypes.
  • Applied machine learning algorithms to interpret Cell Painting data and identify relevant morphological changes.
  • Conducted a high-throughput chemogenomics screen of approximately 5,000 FDA-approved or clinical-stage compounds.

Main Results:

  • Identified several small molecules and target classes with the potential to reverse activated fibrotic phenotypes in intestinal myofibroblasts.
  • Demonstrated the efficacy of integrating morphological analysis with artificial intelligence for disease-relevant cell screening.
  • Validated a novel phenotypic screening platform for accelerating the discovery of intestinal fibrosis therapeutics.

Conclusions:

  • The developed phenotypic screening platform significantly advances the identification of drug targets for intestinal fibrosis.
  • This AI-augmented approach offers a powerful tool for discovering treatments for fibrotic diseases.
  • The identified compounds and target classes represent promising avenues for future therapeutic development in intestinal fibrosis.