Common Variable Immunodeficiency patients with a phenotypic profile of immunosenescence present with thrombocytopenia

Abstract

Common variable immunodeficiency (CVID) is a heterogeneous group of diseases. Our aim was to define sub-groups of CVID patients with similar phenotypes and clinical characteristics. Using eight-color flow cytometry, we analyzed both B- and T-cell phenotypes in a cohort of 88 CVID patients and 48 healthy donors. A hierarchical clustering of probability binning "bins" yielded a separate cluster of 22 CVID patients with an abnormal phenotype. We showed coordinated proportional changes in naïve CD4+ T-cells (decreased), intermediate CD27- CD28+ CD4+ T-cells (increased) and CD21low B-cells (increased) that were stable for over three years. Moreover, the lymphocytes' immunophenotype in this patient cluster exhibited features of profound immunosenescence and chronic activation. Thrombocytopenia was only found in this cluster (36% of cases, manifested as Immune Thrombocytopenia (ITP) or Evans syndrome). Clinical complications more frequently found in these patients include lung fibrosis (in 59% of cases) and bronchiectasis (55%). The degree of severity of these symptoms corresponded to more deviation from normal levels with respect to CD21low B-cells, naïve CD4+ and CD27− CD28+ CD4+ T-cells. Next-generation sequencing did not reveal any common genetic background. We delineate a subgroup of CVID patients with activated and immunosenescent immunophenotype of lymphocytes and distinct set of clinical complications without common genetic background.

Figure 1. Unsupervised clustering of the phenotypic profiles of lymphocytes in CVID patients.

Hierarchical tree of CVID patients (black) and healthy donors (grey) clustered with respect to B- and CD4+ T-cell phenotypes using a probability binning approach. A distinct cluster of CVID AcT patients (red frame) was formed (A). Although the comparison showed a non-random pattern, the composition of the clusters did not follow the EUROclass nor Freiburg classifications (inset A). To determine the phenotypic hallmarks of the clusters, differentially fed bins in CVID-Act were determined (highlighted as colored rectangles) and projected over a typical phenotype of normal healthy controls (black dots) (B). A classification tree was fitted using conventionally gated B-cell and CD4+ T-cell subsets (see gating in Fig. 2) in order to find a simplified definition of CVID- AcT patients (C).

Figure 2. Typical immunophenotype of a CVID-AcT patient.

Gating strategy for B and T-cell analysis (A), with relevant B and CD4+ T cell subsets displayed in healthy donor (B) and in CVID AcT patient (C), gates used for simplified definition are highlighted in red.

Figure 3. Trombocyte counts are significantly reduced in CVID-AcT patients, patients with splenomegaly and decrease with age.

The thrombocyte count is significantly lower in CVID-AcT patients compared to other CVID patients (A), as well as in patients with splenomegaly (B) and it decreases with age (C).

Figure 4. Stability and relationships of defining lymphocyte populations.

CVID-AcT patients’ samples contain the same proportion of the defining populations (naïve, intermediate CD28+CD27− CD4+ T-cells, CD21lo B-cells) when measured more than 3 years apart, which is in contrast to the dynamic compartment of terminal effector CD8+ T-cells. The line connecting the two time points is shown to visualize the trends (A). Investigation of the mutual relationship of the relevant populations shows orchestrated changes in the patients’ immunophenotypes (B).

Figure 5. Clinical signs of CVID-AcT patients are correlated with immunosenescence (activation of immune system and loss of naive T-cells).

The phenotype of CVID-AcT exhibited signs of immunosenescence in agreement with the higher average age of patients in this cluster. The phenotypic hallmarks of the CVID-AcT patients correlated significantly with age (compared to the T-cell hallmarks both in CVID patients and healthy donors). Additionally, the T-cell phenotype seems to change with age at the same rate in the CVID patients and healthy donors (with no significant difference in the slope), but the CVID patients begin with a decreased number of T-cells compared to the healthy controls (A). The phenotypic hallmarks of CVID-AcT correlated significantly (after adjustment for multiple testing) with the severity of their clinical signs (B).

Figure 6. CVID-AcT is characterized by lymphocyte activation, exhaustion and IL-10 serum elevation.

Markers of lymphocyte activation, exhaustion and senescence (A) and serum cytokine levels (B) in healthy controls, other CVID patients and Cluster AcT CVID groups. Asterisk (*) denotes a significant difference between the groups (adjustment p-value using the Benjamini & Yekutieli procedure).