Ex vivo imaging-based high content phenotyping of patients with rheumatoid arthritis.

Background: High content imaging-based functional precision medicine approaches have been developed and successfully applied in the field of haemato-oncology. For rheumatoid arthritis (RA), treatment selection is still based on a trial-and-error principle, and biomarkers for patient stratification and drug response prediction are needed.

Methods: A high content, high throughput microscopy-based phenotyping pipeline for peripheral blood mononuclear cells (PBMCs) was developed, allowing for the quantification of cell type frequencies, cell type specific morphology and intercellular interactions from patients with RA (n = 65) and healthy controls (HC, n = 33). Samples were exposed to a curated set of RA-specific small molecules, biologicals and reference stimuli for 24 h to assess ex vivo drug effects. Data on ex vivo PBMC phenotypes were integrated with information on patients' in vivo medication and disease activity.

Findings: The unbiased data from in total 6.9e8 individual cells were collected and allowed for the identification of PBMC phenotypes specific to disease activity as well as in vivo and ex vivo treatment. The arrayed ex vivo drug perturbation enabled the systematic characterization of drug effects, clustering by mode of action and uncovered morphologic alterations associated with biologic disease-modifying anti-rheumatic drug (DMARD) treatment. Individual in vivo treatment regimens translated into altered immune cell abundances in patients with a comedication of conventional synthetic DMARDs when compared to HCs. Global integration of PBMC characteristics led to clustering of patients according to disease activity and correlation with clinical data.

Interpretation: The application of the developed screening tool demonstrates a technical proof-of-concept for feasibility of a functional precision medicine approach to the ex vivo immunophenotypic characterisation of patients with RA.

Funding: This work was supported by the Austrian Academy of Sciences, the Medical University of Vienna and a grant (RMG2235 to L.X.H.) from the European Alliance of Associations for Rheumatology (EULAR).