Pre-treatment immune profiling by flow cytometry and single-cell transcriptomics identifies distinct features in B and T cells in rituximab responders and non-responders

Abstract

Abstract Introduction Immune thrombocytopenia (ITP) is a heterogeneous autoimmune disorder defined by low platelet counts (<100×10⁹/L) and unclear pathogenesis. While B cell-mediated autoantibodies contribute to platelet destruction, autoreactive T cells are increasingly implicated in ITP patients (pts) beyond B-cell involvement. Rituximab, a CD20-targeting B cell-depleting therapy, is commonly used in ITP, yet only 50–60% of pts respond initially, and most relapse within five years. Biomarkers to predict clinical benefit from available B-cell depleting antibodies are lacking. To identify immune signatures predictive of rituximab response, we performed multi-parameter flow cytometry and single-cell RNA and T cell receptor (TCR) sequencing on pts before treatment. Methods Patients previously treated with rituximab as a second- or later-line treatment with available clinically annotated blood samples were identified from the Imperial College biobank. Complete responders (CR) were defined by platelet counts >100x109/L for ≥6-9 months after ≤12 weeks of treatment; non-responders (NR) had counts <30x109/L for 75% of the time for 6 months after ≥12 weeks of treatment. Partial responders were excluded from this study. Peripheral blood mononuclear cells (PBMCs) were isolated from whole blood and cryopreserved. Deep immunophenotyping of major immune subsets including T, B, and natural killer (NK) cells was performed using 33-colour spectral flow cytometry (Sony ID7000; FlowJo v10.10.0). Combined single-cell (sc) RNA and TCR sequencing was performed on one CR and one NR using the 10x Genomics platform (Chromium Single Cell VDJ, and 5′ Library kits). Cell Ranger was used for sample demultiplexing and unique molecular identifier counting. Gene expression and TCR analyses were conducted using Seurat v5.3 and scRepertoire v2.2.1, respectively. Results 18 pts were recruited; 12 CRs and 6 NRs. Flow cytometry revealed that NRs had significantly heightened baseline immune activation across multiple immune subsets. They exhibited a higher proportion of unswitched memory B cells (p=0.003); and CD95 expression was elevated on plasma cells (p=0.01), CD4+ effector memory T (Tem) (p=0.04), CD8+ Tem (p=0.02), and NK cells (p=0.0009). NRs also showed increased HLA-DR expression on CD8+ Tem (p=0.02) and higher PD-1 levels on CD8+ terminally effector memory T (TEMRA) cells (p=0.02) compared to CRs. ScRNA sequencing revealed distinct immune transcriptional profiles between a CR and a NR. In the CR, B cells were enriched for B cell receptor (BCR) signalling and early activation genes (e.g., CD79A, SYK, TCL1A, FOS, and JUN). Further sub-clustering showed a coordinated functional activation across B cell development with naïve B cells expressing interferon response and early activation signatures commonly seen in vaccine-induced immunity, chemokine- and cytokine-activated unswitched memory cells as well as antibody-producing plasmablasts, indicating a robust, activated humoral profile. In contrast, the NR exhibited reduced BCR activity, impaired B cell differentiation and enrichment of T cell-mediated inflammatory pathways. Sub-clustering of T cells revealed transcriptional features of chronic activation, including upregulated oxidative phosphorylation genes (e.g., NDUFA1, NDUFB8) across all T cell subsets. CD8+ TEMRA cells in the NR showed high inhibitory marker expression (TOX, LAG3, EOMES) and lacked memory genes (IL7R, TCF7, LEF1), consistent with flow cytometry evidence of exhaustion. scTCR sequencing further showed greater clonal expansion (clones occupying > 1% of the TCR repertoire) and reduced TCR diversity in the NR, indicating sustained activation and expansion of T cells. Conclusions Our combined flow cytometry and single-cell transcriptomic and TCR profiling reveal key pre-treatment immune signatures associated with rituximab response in ITP. CRs exhibit an activated B cell profile, while NRs display sustained clonal expansion of highly activated and terminally differentiated T cells. These findings suggest that baseline immune profiling can help identify the dominant immune cell populations contributing to disease in individual patients. Assessment of B cell functionality and T cell clonal expansion may therefore serve as predictive biomarkers for response to CD20-depleting therapies, as well as identify alternative treatment strategies for some patients. A prospective validation in clinical trials is warranted.