Predicting gene-level sensitivity to JAK-STAT signaling perturbation using a mechanistic-to-machine learning...

Area of Science:

• Immunology
• Computational Biology
• Molecular Biology

Background:

• The Janus kinase (JAK)-signal transducer and activator of transcription (STAT) pathway is crucial for integrating cytokine signals.
• Understanding the specificity of STAT transcription factor function is essential for deciphering complex cellular responses.
• Interleukin (IL)-6 and IL-10 signaling utilize common STATs but exhibit distinct temporal dynamics and functional outcomes in macrophages.

Purpose of the Study:

• To develop a computational framework for predicting cytokine-induced gene expression from STAT phosphorylation dynamics.
• To model macrophage responses to IL-6 and IL-10, focusing on STAT pathway signaling.
• To link STAT signaling dynamics to specific gene expression patterns and assess the impact of JAK2 inhibition.

Main Methods:

• Developed a mechanistic-to-machine learning computational model.
• Modeled STAT phosphorylation dynamics in response to IL-6 and IL-10 in macrophages.
• Predicted and validated cytokine-specific gene expression changes and JAK2 inhibition effects.

Main Results:

• Identified cytokine-specific genes associated with late pSTAT3 (phosphorylated STAT3) time frames.
• Observed a preferential reduction in pSTAT1 (phosphorylated STAT1) upon JAK2 inhibition.
• Predicted and validated genes sensitive or insensitive to JAK2 variation, linking STAT dynamics to gene expression.

Conclusions:

• Successfully linked STAT signaling dynamics to gene expression, providing a foundation for future research.
• The developed model can aid in targeting pathology-associated STAT-driven gene sets.
• This work represents a first step towards multi-level prediction models for understanding and perturbing signaling system outputs.