Abstract
Introduction. Immune thrombocytopenia (ITP) is a rare, acquired autoimmune disorder characterized by isolated thrombocytopenia with a risk of potential life-threatening bleeding. Effective treatment is therefore critical to patient outcomes. While first-line therapies can quickly restore platelet counts, they do not modify the course of the disease with around 70% of adults evolving to chronic disease and requiring further treatments. Second-line treatments are mostly based on thrombopoietin receptor agonists and rituximab, an anti-CD20 monoclonal antibody. For the latter, response rates are around 60% at one year and 30% at five years, with no reliable predictive clinical or biological markers identified to date.
In the aim to identify biological predictive factor of response to rituximab, we performed a translational research study to characterize the immune profile of ITP patients before treatment.
Methods. This study was conducted using a biobank of cryopreserved peripheral blood mononuclear cells (PBMC) samples from adult patients with ITP, prospectively enrolled in clinical trials led by the French National Referral Center for Autoimmune Cytopenias (University hospital of Dijon, France) between 2010 and 2020 (ClinicalTrials.gov: NCT02821572 and NCT02042560). The diagnosis of ITP was established according to international criteria. All PBMC used here were collected prior to rituximab administration. To improve the robustness of our study, long-term response at 5 years was considered based on platelet count (complete response: >100 ×10⁹/L; partial response: 30-100×10⁹/L with at least a twofold increase from baseline) without concomitant treatment. Platelet count <30×10⁹/L or need for additional treatment 8 weeks after rituximab was considered as non-response. Based on these criteria, 20 responders and 18 non-responders were identified.In the first phase, single-cell RNA sequencing (scRNA-seq) was performed using PBMC from 4 responders and 4 non-responders (matched for age, sex, ITP duration and prior therapies), on an Illumina NextSeq2000 platform. This unbiased approach aimed to explore the immunopathogenic mechanisms and to identify potential predictive biomarkers of response.
To validate these findings, comprehensive immunophenotyping of circulating immune cells was performed using flow cytometry (Attune, ThermoFisher) on the entire cohort.
Results. In non-responders, scRNA-seq analyses revealed a more advanced differentiation state across all immune cell populations. B cells exhibited a more mature phenotype, with higher proportions of both unswitched and switched memory B cells. CD8 T cells also showed a memory-differentiated profile, including central memory, effector memory and effector memory T cells re-expressing CD45RA (EMRA) subsets, with increased expression of cytotoxicity markers. In addition, monocytes presented a pro-inflammatory transcriptional profile in these patients.
These findings were validated by flow cytometry, demonstrating a significantly higher proportion of switched memory B cells (CD19⁺CD27⁺IgD⁻; 25.8% [18.7-34.7] vs. 13.4% [7.9-25.1], p=0.01) and a lower naive/switched memory B cell ratio (1.8 [0.9-2.9] vs. 4.4 [1.5-6.6], p=0.02) in non-responders. Preliminary results concerning CD8 T cells and based on data from 13 responders and 11 non responders, showed an increased frequency of CD8 TEMRA producing IFN-γ at baseline (without stimulation) in non-responder patients (1.5% [1-1.7] vs. 0.7% [0.6-1.1] of total CD8 TEMRA cells, p=0.03).
Conclusions. Our findings highlight distinct immunological profiles between long-term responder and non responder ITP patients to rituximab. Non-responders exhibited more differentiated B cells and more differentiated and activated CD8 T cells, that may contribute to rituximab failure.
