Alterations in the Immune Phenotype of Adaptive Immune Cells in Sickle Cell Disease: A Conventional and Computational Flow Cytometry Analysis

Rationale

Sickle cell disease (SCD) is recognized as a chronic inflammatory condition. Impaired vaccination responses, frequent alloimmunization, a high prevalence of autoimmune diseases and an increased risk of graft rejection following hematopoietic stem cell transplantation indicate altered adaptive immune responses. Therefore, elucidating the phenotype of peripheral blood mononuclear cells (PBMC) in SCD is essential for improving insight and future patient management.

Aim

To evaluate the characteristics of adaptive immune cells involved in chronic inflammation and related pathophysiology in SCD patients, we investigated the differences in PBMC subsets between SCD patients and ethnicity- and age-matched healthy controls (HCs).

Methods

PBMCs from HbSS/HbSβ⁰ patients (≥ 12 years) in steady state and HCs were analyzed using a 5x14-color flow cytometry panel for in-depth characterization of B, T, T-helper (Th), T follicular-helper (Tfh), natural killer (NK) cells, monocytes and dendritic cells (DCs). Measurements were standardized using rainbow calibration beads on a Symphony flow cytometer. Cell populations were identified manually and computationally by conventional gating (FlowJo) and unsupervised clustering (FlowSOM), respectively. Statistics included Benjamini-Hochberg correction of false discovery rate.

Results

The study included 46 patients (37% female, 94% HbSS, 78% on hydroxyurea): 26 adults (median age 25 years, range 18 - 57) and 20 adolescents (16 years, 12 - 18). Among 18 HCs (33% HbAS), 12 were adults (30 years, 22 - 42) and 6 were adolescents (16.5 years, 13 - 18). Subgroup analysis revealed no differences in proportions of immune cell subsets between age groups or hydroxyurea use.

Gated analysis of B cells (CD45+CD14-CD19+) showed major differences between patients and HCs: patients had higher naive transitional B cells (IgD+CD27-, p <0.001) and lower proportions of unswitched memory (IgD+CD27+, p <0.001), switched memory (IgD-CD27+, p <0.001), and double negative memory (IgD-CD27-, p= 0.003) B cells. Patients also had fewer activated naive/activated memory (CD24-CD38-, p= 0.029) B cells within total B cells. Computational analysis confirmed these results with 11 of 18 FlowSOM metaclusters showing significant differences between patients and HCs: one metacluster, representing naive transitional B cells and containing 68% of total B cells, was enriched in patients (p= 0.004) but with a lower mean fluorescence intensity (MFI) of CD21 (p= 0.007). Several metaclusters containing memory B cell subsets were less abundant in patients.

Gated analysis of T cells (CD45+CD14-CD3+) revealed no differences in CD4+ or CD8+ cells, regulatory (CD4+CD25+CD127-FoxP3+), naive (CCR7+CD45RA+), effector (CCR7-CD45RA+), central memory (CCR7+CD45RA-), effector memory (CCR7-CD45RA-) T cells, or Th subsets. However, patients had more γδ+ T cells (p <0.001) and less αβ+ T cells (p <0.001) compared to HCs. Computational analysis showed that MCs containing CD4-CD8- double negative (DN) T cells were significantly enriched in patients compared to HCs (p= 0.008), which was confirmed by manual gating of DN T cells.

SCD patients had fewer conventional CD11c+CD141+ cDC1s (p <0.001) and CD11c+CD1c+ cDC2s (p= 0.021), but more CD123+ plasmacytoid DCs (p= 0.024). Total NK cells were decreased, but with a higher proportion of regulatory CD56++ NK cells (p= 0.005). Patients had more classical monocytes (CD14+CD16-, p= 0.001), but fewer Slan+ and non-classical monocytes (CD14-CD16+, p <0.001). FlowSOM analysis revealed, that metaclusters containing CD56-, but Slan+ and/or CD16+ monocytes, one metacluster matching the phenotype of cDC1s, and one metacluster containing CD11c+ NK cells were significantly diminished in patients.

Conclusion

We present a comprehensive and unique computational immune phenotype evaluation of PBMCs in SCD patients. Our analysis reveals significant perturbations in B cell subsets, with a shift towards naive transitional and less memory B cells, likely due to SCD-related hyposplenia. SCD patients also exhibit increased DN and γδ+ T cells, common in autoimmune and chronic inflammatory conditions. Finally, we observed a decrease of cDCs, Slan+ monocytes and NK cells expressing CD11c, crucial for regulating immune responses. Further study of these PBMC subsets may reveal new insights into chronic inflammation and adaptive immune dysregulation in SCD.