Mapping the Spatial Proteome of Individual Leukemia Cells Undergoing Fludarabine Treatment

Introduction: Fludarabine, a purine nucleotide analogue, is integral to preparative regimens for CAR-T cell therapy and allo-SCT in treating B-cell pediatric acute lymphoblastic leukemia (B-ALL). By disrupting DNA synthesis and repair, fludarabine induces apoptosis in proliferating leukemia cells. It also plays a vital role in immunosuppression, lymphodepletion, and tumor burden reduction, optimizing the efficacy of subsequent therapies. Despite its critical role, the molecular impact of fludarabine on leukemia cells and its efficacy factors remain inadequately understood.

Methods: We utilized a cutting-edge single-cell spatial proteomics assay, Molecular Pixelation (MPX, Karlsson et al. Nat Methods 2024), an optics-free DNA sequencing-based method to analyze the cell surface interactome in response to fludarabine. The glucocorticoid-resistant ETV6::RUNX1 cell line REH and the BCR::ABL1 cell line SUP-B15 were incubated with and without fludarabine for 72 hours at varying concentrations, with repeated technical repetitions. Approximately 1,000 surviving single cells per condition were analyzed using the MPX assay (Pixelgen Technologies), wherein antibody–oligonucleotide conjugates generated nanometer-sized molecular pixels, read out by sequencing on a NovaSeq X Plus instrument (Illumina). Spatial proteomics networks for 76 proteins were computationally reconstructed from DNA-sequencing reads by inferring the relative locations of antibody–oligonucleotide conjugates, creating ~1,000 spatially connected neighborhoods per cell. Orthogonal immunocytochemistry (ICC) data were generated to validate the top protein interactions.

Results: We have previously shown that ex-vivo drug screening in combination with single-cell transcriptome sequencing (scRNA-seq) readout can elucidate transcriptional effects of fludarabine (Gezelius et al. NAR Genomics and Bioinformatics 2024). This approach confirmed fludarabine’s disruption of DNA synthesis and repair in dividing cells, and in addition revealed previously unknown extensive dysregulation of transcripts encoding cell surface proteins. The scRNA-seq data do not adequately resolve the spatial nor functional dynamics of the cell surface proteome, leaving this important aspect of response to fludarabine unexplored. In this study, we take advantage of recent advancements in single-cell spatial proteomics profiling to spatially elucidate how the cell surface proteome organization and function is modulated during fludarabine treatment.

Spatial statistical analysis revealed novel patterns in cell surface organization during fludarabine treatment. The MPX assay identified 27 proteins with significantly altered abundance post-treatment (adj. Wilcoxon p < 0.01), consistent with the increased abundance of cell-surface proteins observed in our previous scRNA-seq data. Additionally, 26 proteins exhibited changes in polarity (adj. Wilcoxon p < 0.01) after fludarabine treatment, with most proteins shifting from a dispersed to a polarized state. We also detected 138 protein pairs with modified co-localization patterns. Notably, fludarabine enhanced the abundance, polarization, and co-localization of cadherin proteins CD82 and CD53, suggesting a potential therapeutic scaffold. This finding was further validated through orthogonal ICC experiments.

Conclusions: This study advances our understanding of the spatial organization of the cell surface proteome in leukemia cells following fludarabine exposure. The findings provide new insights into fludarabine's effects at the single-cell level and may inform the development of more effective targeted therapies.