Session 1: Artificial Intelligence & Cancer

Dynamical modelling of gene regulatory networks in Chronic Lymphocytic Leukaemia: Understanding the role of immune cell interactions in cancer cell phenotypes

Purpose: In cancer systems, the tumour microenvironment (TME) can be seen as interacting cells that collectively shape cellular behaviours and responses to internal and external stimuli. Although much is known about immune cell reprogramming, the detailed molecular insights of cancer cell behaviour remain incomplete. To bridge this gap, we have focused on leveraging data-driven approaches to build dynamic models of gene regulation, studying the functional processes determining the cellular interactions at the molecular level and investigating how these interactions characterise cancer cell phenotypes.



Methods: Using time-course RNAseq datasets from in vitro cultures of Chronic Lymphocytic Leukaemia (CLL) patient samples, we perform gene regulatory network (GRN) inference [1], transcription factor activity (TF) analysis, and independent component analysis (ICA) [2], to highlight important gene modules and their functions in CLL cells. Furthermore, we build a dynamical model of CLL cells [3], thus better understanding their behaviour and response to environmental stimuli. Utilising two experimental conditions (CLL PBMC culture, and CLL cell monoculture), we investigate the fundamental role of immune cells present in the culture in CLL cell viability, transcriptomic state and likelihood of expressing proliferation signatures.



Results: Our results on ICA highlight gene (TF) groups that may correspond to distinct functional modules or regulatory circuits within the CLL transcriptional landscape, indicating that CLL cells may depend on discrete, modular regulatory programs to adapt to varying conditions, such as the presence of immune cells. Furthermore, in silico KO experiments on ligands involved in interaction with immune cells, showed phenotypic changes in CLL cells, particularly in regulatory pathways related to proliferation. 



Conclusions: Our study highlights the potential of dynamic GRN models to capture the complex transcriptional programs in CLL and the importance of immune cell interactions in modulating CLL cell phenotypes. By identifying key regulatory modules and critical pathways influenced by immune components, these findings pave the way for targeted therapeutic strategies that consider the broader cellular context within the TME.



References

[1] Huynh-Thu and Geurts, 2018.

[2] Captier et al, 2022.

[3] Paulevé, 2023.