Frequency and Expression of Inhibitory Markers of CD4+CD25+FOXP3+ Regulatory T Cells in Patients with Common Variable Immunodeficiency

Authors


Correspondence to: Asghar Aghamohammadi, Children's Medical Center Hospital, 62 Qarib St., Keshavarz Blvd., Tehran 14194, Iran. E-mail: aghamohammadi@tums.ac.ir

Abstract

Common variable immunodeficiency (CVID) is the most symptomatic primary antibody deficiency associated with recurrent infections and chronic inflammatory diseases as well as autoimmunity. CD4+CD25+FOXP3+ regulatory T cells (Tregs) are critical T cell subsets for maintaining self-tolerance and regulation of immune response to antigens thus play a pivotal role in preventing autoimmunity. Thirty-seven CVID patients and 18 age-/sex-matched controls were enrolled. Peripheral blood mononuclear cells (PBMCs) were obtained from both groups, and the percentage of Tregs was calculated using flow cytometry method. The mRNA expression of Tregs' surface markers cytotoxic T lymphocyte–associated antigen-4 (CTLA-4) and glucocorticoid-induced tumour necrosis factor receptor (GITR), which are associated with Tregs' inhibitory function, was compared between patients and controls by quantitative real-time PCR TaqMan method. The results revealed that the frequency of Tregs was significantly lower in CVID patients than normal individuals (P < 0.001). In addition, CVID patients with autoimmunity were found to have markedly reduced proportion of Tregs compared to those cases without autoimmune diseases (P = 0.023). A significant difference was seen in factor forkhead box P3 (FOXP3) expression between CVID patients and controls (P < 0.001). The mRNAs of CTLA-4 and GITR genes were expressed at lower levels in CVID patients compared to control group (P = 0.005 and <0.001, respectively). Our findings showed reduced proportion of Tregs in CVID patients together with downregulation of FOXP3 protein and diminished expression of inhibitory Tregs' markers. It could be concluded that all of these changes may be responsible for cellular immune dysregulation observed in these patients especially those with autoimmune manifestation.

Introduction

Common variable immunodeficiency (CVID) is a heterogeneous group of disorders characterized by hypogammaglobulinaemia, defective specific antibody production and an increased susceptibility to recurrent and chronic infections [1-3]. Patients with CVID also have an increased incidence of autoimmune disorders and cancers [4-6]. In addition to reduced Ig production by B cells, several defects in T cell response have been reported in CVID patients including impaired cell proliferation and cytokine production as well as reduced T cell numbers [7].

The CD4+CD25+FOXP3+ regulatory T lymphocytes (Tregs) constitute about 5–10% of the peripheral blood CD4+ T cells and have an indispensable role in maintaining self-tolerance and immune response to self and non-self antigens [8, 9]. This unique subset of CD4+ T cells have been implicated in regulating immune response in different conditions like allergic diseases, malignancy, graft vs. host diseases as well as autoimmune disorders [9, 10]. Although cell to cell contact has been considered the major mechanism of Tregs-mediated suppression, the production of regulatory cytokines like Il-10, IL-35 and TGF-β by Tregs should also be noted [8-10].

There are increasing evidences indicating the reduced frequency of Tregs in autoimmune diseases, which has been shown to have inverse correlation with clinical parameters [11-16]. Recently, few reports have been published indicating reduced numbers of Tregs in CVID patients and its correlation with chronic inflammation, splenomegaly and autoimmune manifestation in these patients [17-21]. In this study, we proposed to investigate Tregs' frequency and transcription FOXP3 protein expression in CVID patients. We also evaluated for the first time the mRNA expression of surface markers CTLA-4 and GITR, which are associated with the inhibitory functions of Tregs in CVID patients and compared the results with healthy controls.

Materials and methods

Subjects

Thirty-seven patients with CVID who were referred to division of clinical immunology and allergy at Children's Medical Center of Tehran University of Medical Sciences were enrolled in this study. The diagnosis of CVID disease was based on defined criteria by PAGID (Pan-American Group for Immunodeficiency) and ESID (European Society for Immunodeficiencies) [2]. All patients were receiving monthly regular intravenous immunoglobulin replacement therapy. Medical history and clinical phenotypes of CVID patients were given from the national primary immunodeficiency registry [1, 22, 23].

Eighteen sex- and age-matched healthy volunteers who have no history of autoimmune disease, malignancy and/or any immunodeficiency were chosen as control group. The study was approved by the ethics committee of Tehran University of Medical Sciences, and written informed consent was obtained from all participants before sampling.

Immunological laboratory assay

Complete blood count was evaluated by the cell counter and Westergren method, using anticoagulated whole blood, respectively. Serum levels of IgG, IgA and IgM were measured by immunoturbidimetry (Behring Nephelometer, Behringwerke, Marburg, Germany), and lymphocyte subpopulations of CD3, CD4, CD8 and CD19 were counted by flow cytometry (Partec PAS, Münster, Germany) at the time of study. Immunoglobulin E and antibody responses against diphtheria were measured, using an enzyme-linked immunosorbent assay (ELISA).

Isolation of peripheral blood mononuclear cells

The blood samples were collected in ethylenediaminetetraacetic acid (EDTA) containing tubes. Peripheral blood mononuclear cells (PBMCs) were obtained from both patients and controls using Ficoll-Paque (Lymphoflot, Bio-Rad, Germany) density gradient centrifugation. Cells were washed once with RPMI 1640 (Sigma, Germany) and prepared for surface staining.

Immunofluorescent staining of the cells and flow cytometric analysis

For surface staining, 1 × 10cells were resuspended in 100 μl flow cytometry staining buffer (eBioscience, San Diego, CA, USA). Cells were incubated with fluorescein isothiocyanate (FITC)-labelled anti-CD4 (clone RPA-T4, eBioscience) and phycoerythrin (PE)-labelled anti-CD25 (clone BC96, eBioscience) antibodies for 30 min at 4 °C in the dark. For intracellular staining, after permeabilization with fixation/permeabilization buffer (eBioscience), PE-/Cy5-labelled anti-FOXP3 antibody (clone PCH101, eBioscience) was added and incubated for 30 min at 4 °C in the dark. FITC- and PE-conjugated mouse IgG1 and PE-/Cy5-conjugated rat IgG2a antibodies were used as the isotype control antibodies.

Quantitative Real-Time PCR analysis

Total RNA was extracted from CD4+ T cells using QIAzol lysis reagent (Qiagen GmbH, Hilden, Germany) followed by cDNA synthesis with M-MuLV reverse transcriptase enzyme (Fermentas Life Science, EU). Quantitative real-time PCR was performed using TaqMan Premix Ex Taq™ (Perfect Real-Time) master mix (Takara, Japan). The PCR primer pairs and probes were as follows: CTLA-4, 5′-CATGGACACGGGACTCTACAT-3′, 5′-GCACGGTTCTGGATCAAT TACATA-3′ and 5′-FAM-TGCAAGGTGGAGCTCATGTACCCACC-TAMRA-3′, GITR, 5′-TGCAAACCTTGGACAGACTGC-3′, 5′-ACAGCGTTGTGGGTCTTGTTC-3′ and 5′-FAM-CCAGTT CGGGTTTCTCACTGTGTTCC-TAMRA-3′. For increasing the validation of our test, two housekeeping genes were selected: TBP (TATA-binding protein) and YWHAZ (a signal transducer molecule that binds to phosphoserine-containing proteins) in which their primer and probe sequences were 5′-TTCGGAGAGTTCTGGGATTGTA-3′, 5′-TGGACGTTCTTCA CTCTTGGC-3′ and 5′-FAM-CCGTGGTT CGTG GCTCTCTTATCCTCA-TAMRA-3′ for TBP and 5′-AAGTTCTTGATCCCCAATGCTT-3′, 5′-GTCTGATAGG ATGTGTTGGTTGC-3′ and 5′-FAM-TATGCTTGTTGTGACTGATCGACAATCCC-TAMRA-3′ for YWHAZ genes. The mRNA was quantified with ABI 7500 software (Applied Biosystems) in duplicate wells, and the Ct values for target and housekeeping genes were calculated in both patients and controls. The efficacy of our test was 1, which was obtained by serial dilution of both target and housekeeping genes. The relative expression of CTLA-4 and GITR genes in CVID patients vs. controls was analysed with REST 2009 software, and expression levels were normalized to both TBP and YWHAZ housekeeping genes.

Statistical analysis

Comparison between patients and healthy controls was carried out with Student's t-test, and for more than two groups, anova test was used. For correlation analysis, Pearson correlation test was performed. P-values less than 0.05 were considered significant.

Results

In this study, 37 CVID patients (29 males and eight females) with mean age of 18.6 ± 10.2 years were enrolled. The mean of delay in diagnosis of patients was 5.7 ± 5.4 years. Totally, all patients were followed up for 278 years (7.5 years per patient) receiving monthly regular intravenous immunoglobulin replacement therapy. Twenty-nine of the CVID patients (78.4%) had early onset of disease, and parental consanguinity was documented in 21 cases (56.8%). Among 37 studied patients, autoimmunity phenotype was the most frequent manifestation which recorded in 16 (43.2%) cases. Within them, 7 (43.7%) patients had autoimmune cytopenia (AIHA, ITP and AN) and the remaining nine patients (56.3%) had other type of autoimmunity (hypothyroidism, JRA, systemic lupus erythematosus, psoriasis and autoimmune hepatitis). Other clinical phenotypes and immunological characteristics of patients are illustrated in Table 1.

Table 1. Clinical and immunological characteristics of 37 common variable immunodeficiency patients and 18 healthy controls
SubjectsHealthy controls (n = 18)Total patients (n = 37)Patients with reduced Tregs (n = 12)Patients with normal Tregs (n = 25)P-value
Mean age (range)17.9 ± 9.218.6 ± 10.216.6 ± 8.119.6 ± 11.10.42
Sex (male/female)14/429/88/421/40.23
Onset of disease (early/late)29/89/320/50.52
Parental consanguinity (%)21 (56.8)8/413/120.60
Infectious only phenotype (%)10 (27)1 (8.3)9 (36)0.08
Autoimmunity phenotype (%)16 (43.2)9 (75)8 (32)0.05
Poly-lymphocytic infiltrative phenotype (%)16 (43.2)6 (50)10 (40)0.41
Enteropathy phenotype (%)4 (10.8)3 (25)1 (4)0.09
Malignancy phenotype (%)2 (5.4)2 (16.6)00.45
Immunoglobulin G (mg/ml)1803.1 ± 562.0199.8 ± 144.9239.9 ± 101.0178.9 ± 161.30.08
Immunoglobulin M (mg/ml)95.1 ± 74.228.0 ± 22.426.3 ± 24.728.8 ± 21.50.84
Immunoglobulin A (mg/ml)45.9 ± 18.619.0 ± 28.620.0 ± 16.018.8 ± 34.00.28
Absolute B cell count (/ml)427.2 ± 93.0377.9 ± 47.3232.9 ± 35.4433.1 ± 50.80.31
Absolute T cell count (/ml)3799.3 ± 2824.02302.4 ± 1352.81625.5 ± 1204.42593.5 ± 1333.80.07
Absolute helper T cell count (/ml)2550.2 ± 1712.9820.5 ± 486.8639.0 ± 591.2898.3 ± 427.70.18
Absolute cytotoxic T cell count (/ml)1609.0 ± 1055.21338.6 ± 1137.4780.2 ± 497.71589.9 ± 1260.20.02
CD4/CD8 ratio1.52 ± 0.70.78 ± 0.490.77 ± 0.480.78 ± 0.500.95
CD4+CD25+FOXP3+ Tregs in CD4 population3.57 ± 1.071.81 ± 0.721.0 ± 0.282.18 ± 0.55<0.001
FOXP3 expression in PBMC3.83 ± 0.982.91 ± 0.52.66 ± 0.453.1 ± 0.540.021

Frequency of CD4+CD25+FOXP3+ Tregs in patients and controls

Flow cytometry was carried out using a Partec flow cytometer (Partec PAS, Germany), and lymphocytes were gated based on their forward and side scatter. The population of Tregs was obtained by calculating the percentage of CD25+ FOXP3+ double-positive cells within CD4+ gate (Fig. 1). Data were analysed with FlowMax software (Partec PAS, Germany).

Figure 1.

Gating strategy for defining the Tregs' frequency in a patient with reduced Tregs (A) compared to a representative healthy control (B). First, lymphocytes were gated based on their forward and side scatters. The percentage of Tregs was measured by calculating the percentage of CD25+ FOXP3+ double-positive cells within CD4+ population using FlowMax software. NC, negative control.

Analysis of our results showed that the frequency of Tregs was significantly lower in CVID patients than normal individuals (1.81 ± 0.72 vs. 3.57 ± 1.07; P < 0.001, Fig. 2). Based on the two standard deviation below the mean of Treg cells in normal group, the cut-off point was defined as 1.43% and those had count lower than this point were considered to have reduced Tregs. The percentage of CD4+CD25+FOXP3+ Tregs for a patient with reduced Tregs (1.01%) is represented in Fig. 1 compared with a normal individual (5.6%).

Figure 2.

Comparison of the frequency of CD4+CD25+FOXP3+ Tregs (among CD4+ population) in CVID patients, healthy controls and CVID subgroups; +AID = patients with autoimmune disease, −AID = those without autoimmunity.

Furthermore, FOXP3 protein expression was analysed based on the FOXP3 mean fluorescence intensity (MFI) in PBMCs. As shown in Fig. 3, FOXP3 protein was decreased in CVID patients than controls (2.91 ± 0.52 vs. 3.83 ± 0.98, P < 0.001). A positive correlation was seen between the frequency of Tregs and FOXP3 expression (r = 0.42, P = 0.01). The suppression assay was performed in the ratio 1:1 Treg/Tres. The percent of suppression was calculated in CVID patients and healthy controls as the indicator of Tregs' inhibitory function. The Tregs' suppressor capacity is markedly diminished (two-fold) in CVID patients compared to controls (P < 0.001).

Figure 3.

FOXP3 expression based on the FOXP3 mean fluorescence intensity (MFI) in PBMCs in CVID patients, healthy controls and CVID subgroups; +AID = patients with autoimmune disease, −AID = those without autoimmunity.

Tregs in different clinical and immunological groups of CVID patients

There was no association between gender, parental consanguinity and onset of disease with the level of Tregs. When CVID patients were classified based on the clinical phenotypes, it was observed that the CVID patients with autoimmunity had markedly reduced proportions of CD4+CD25+FOXP3+ Tregs compared to those with infectious only (post hoc analysis; P = 0.035) and those with poly-lymphocytic infiltrative phenotype (post hoc analysis; P = 0.022). Patients with autoimmune diseases also had significant reduction in Tregs compared to the rest of CVID patients without autoimmunity (1.50 ± 0.64 vs. 2.04 ± 0.70, P = 0.023; Table 2). Moreover, CVID patients with autoimmunity had significantly lower expression of FOXP3 protein than those without autoimmunity (2.64 ± 0.39 vs. 3.15 ± 0.52, P = 0.002). The expression of FOXP3 protein in patients with autoimmune cytopenia was 2.43 ± 0.23, which was significantly lower than CVID cases with other types of autoimmunity (3.0 ± 0.58; P = 0.025).

Table 2. Clinical and immunological characteristics of CVID patients with and without autoimmunity
SubjectsPatients with autoimmunity (n = 16)Patients without autoimmunity (n = 21)P-value
Mean age (range)17.2 ± 10.419.7 ± 10.20.47
Sex (male/female)13/316/50.51
Onset of disease (early/late)13/316/50.51
Parental consanguinity (%)10/611/100.35
Immunoglobulin G (mg/ml)255.1 ± 144.0158.3 ± 134.40.49
Immunoglobulin M (mg/ml)30.9 ± 28.625.8 ± 16.70.51
Immunoglobulin A (mg/ml)24.3 ± 13.815.0 ± 8.30.35
Absolute B cell count (/ml)1347.4 ± 135.65265.0 ± 545.30.02
Absolute T cell count (/ml)2148.3 ± 162.52405.1 ± 1142.70.61
Absolute helper T cell count (/ml)563.9 ± 361.0991.6 ± 492.40.016
Absolute cytotoxic T cell count (/ml)1477.9 ± 170.61253.5 ± 630.70.61
CD4/CD8 ratio0.66 ± 0.510.84 ± 0.470.33
CD4+CD25+FOXP3+ Tregs in CD4 population1.50 ± 0.642.04 ± 0.700.023
FOXP3 expression in PBMC2.64 ± 0.393.15 ± 0.520.002

Regression analysis of immunological data of cases failed to show any correlation with level of Tregs; however, the reverse association between serum level of IgG and Tregs was observed in CVID patients (r = −0.36, P = 0.031).

According to the Tregs' cut-off point, 12 CVID patients had reduced number of these cells. These Treg-low patients had meaningfully lower absolute counts of cytotoxic T cells (780.2 ± 497.7 cell/ml) compared to other CVID patients (1589.9 ± 1260.2 cell/ml, P = 0.02). Consistent with previous results, these twelve selected cases had significant different autoimmune manifestation compared to remaining patients (75% vs. 32%, P = 0.05, Table 1).

Expression of CTLA-4 and GITR mRNA in CVID patients

The results revealed that there was a significant reduction in mRNA expression of both CTLA-4 (3.8-fold) and GITR (3.7-fold) genes in CVID patients compared to the control group (P = 0.005 and P < 0.001) (Fig. 4). Moreover, the relative expression of these genes was analysed in CVID patients with autoimmune diseases vs. those without autoimmunity. No difference was observed in relative expression of both CTLA-4 and GITR genes within this subgroup of CVID patients (P = 0.82 and P = 0.23).

Figure 4.

Relative expression of CTLA-4 and GITR genes in CVID patients (the expression levels are represented in logarithmic scale and normalized to healthy controls).

The expression of both genes had no difference between CVID cases with reduced number of Tregs and those with normal Tregs (P = 0.70 for CTLA-4, P = 0.40 for GITR) and between autoimmune CVID cases with autoimmune cytopenia and other types of autoimmunity (P = 0.62 for CTLA-4, P = 0.77 for GITR).

Finally, we assessed any correlation existed between Tregs' frequency and mRNA gene expression of their inhibitory markers: CTLA-4 and GITR in CVID patients and also among CVID subgroups. There was no significant correlation between the frequency of Tregs and expression of both CTLA-4 gene (r = 0.078, P = 0.53) and GITR gene (r = 0.18, P = 0.15) in any of the groups.

Discussion

In the present study, the proportion of the Tregs was investigated in CVID patients to determine whether changes in Tregs' number might be relevant to immune dysregulation observed in these patients. Based on the findings of this study, the proportion of CD4+CD25+FOXP3+ Tregs is markedly reduced in CVID patients than controls. Furthermore, patients with autoimmune diseases have lower percentage of Tregs compared to those without autoimmunity. In agreement with these results, previous studies showed that the frequency of Tregs is decreased in CVID patients and its correlations with chronic inflammation, splenomegaly and autoimmune manifestation have also been described [17-21].

Tregs were initially introduced by Shimon Sakaguchi and his colleagues [24] as a unique subset of CD4+ T cells that constitutively express high levels of surface IL-2 receptor α chain, CD25 and transcription factor FOXP3 and have potent immunoregulatory properties [9, 25]. This population of T lymphocytes also express other markers including CTLA-4, GITR, LAG-3 (CD223), galectin-1 and low levels of CD127 (IL-7 receptor α) [10].

Controlling the homoeostasis of Tregs can be exerted in different aspects like their thymic development and differentiation, half-life in circulation and their tissue redistribution [26]. Therefore, it is tempting to believe that changes in each of these checkpoints might reflect Tregs' populations in peripheral blood of CVID patients particularly those with autoimmune diseases. One possible explanation is the homing of Tregs from blood into the site of inflammation. Defect in thymic development should also be considered because defect in thymopoiesis has been reported in some studies in CVID patients [27, 28].

Common variable immunodeficiency shares many clinical phenotypes with selective IgA deficiency (SIgAD) associating with severe complication, and progression from SIgAD to CVID has also been reported in several cases [29, 30]. In our previous report, it was presented for the first time that the frequency of Tregs is lower in patients with SIgAD, especially those with autoimmune diseases [31]. Therefore, it could be hypothesized that reduced number of Tregs' cells may play a similar role in the pathogenesis of both diseases. Carter et al. [32] conducted a study to compare the levels of regulatory T cells and the activation markers of T cell subsets in 23 CVID patients and to clarify their possible interaction leading to autoimmunity. Similar to finding of this study, they showed that patients especially those with autoimmune manifestation had reduced levels of Tregs compared with control group. Moreover, they found that elevated T cell expression of granzyme B and HLA-DR had another indicators predisposing CVID patients to autoimmunity.

We further investigate the key molecules involved in Tregs' functions including FOXP3, CTLA-4 and GITR markers. In complete agreement with other published data, CVID patients had diminished expression of FOXP3 protein compared to controls as well as those with autoimmunity compared to non-autoimmune ones [18, 20]. Additionally, a positive correlation was seen between the frequency of Tregs and FOXP3 expression. It has been well accepted that stable expression of FOXP3 is necessary for maintaining the development and regulatory characteristics of natural Tregs [33]. Interestingly, there is some evidence describing the conversion of murine CD4+CD25+FOXP3+ Treg cells into CD4+CD25+FOXP3- T cells as a result of FOXP3 downregulation, thus subverting Tregs to T effector and predisposing autoimmunity [34, 35]. Indeed, chronic inflammation seen in CVID disease might create a milieu in which activation of effector T cells may cause downregulation of FOXP3 via production of inflammatory cytokines, thus alter Tregs' proportions and consequently increase the risk of autoimmunity [17]. However, more studies are needed to support this idea.

Our findings in this study indicate that both CTLA-4 and GITR mRNA levels are decreased in CVID patients compared to the control group. This is the first time that CTLA-4 and GITR genes are evaluated at mRNA level in CVID patients. Only one study by Yu et al. showed that the GITR molecule expression is attenuated at protein level (using MFI by flow cytometric analysis) in CD4+CD25highCD127low Tregs from CVID patients with autoimmunity comparing those without autoimmunity and also healthy controls [21].

Several mechanisms for Tregs-mediated immune suppression have been described in which both surface markers (e.g. CTLA-4, GITR, LAG-3) and soluble cytokines (e.g. IL-10, TGF-β and IL-35) have been implicated [8-10]. However, the role of soluble factors is still controversial and cell–cell contact has also been considered as a major aetiology [8-10]. The CTLA-4 and GITR molecules are constitutively expressed at high levels on Tregs' surfaces. The main role of CTLA-4 molecule is to compete with CD28 molecule for CD80/CD86 markers on dendritic cells (DCs) and thus restraining the effector T cell activation [8, 36]. Negative signal transduction of Tregs by CTLA-4 to DCs can convert them to tolerogenic DCs [37]. During the effector phase of an immune response, the GITR molecule promotes Tregs' activation and proliferation, which restrict uncontrolled immune cell activation [38, 39].

Hence, it is possible that changes in CTLA-4 and GITR expression together with downregulation of FOXP3 protein might account for Tregs' dysfunction observed in CVID patients. It is possible that ICOS has the same costimulatory role in Treg activation (like conventional T cells) and genetic defect in ICOS gene has been reported to be associated with susceptibility to CVID and defective Treg function [40]. Therefore, evaluating the expression of ICOS might provide additional data in pathogenesis of CVID and should be considered in future studies. Furthermore, recent study reported that Th17 populations differentiated in vitro from natural naive FOXP3+ Tregs, which should be investigated in another study via evaluation of IL-17-producing cells in CVID patients [41]. However, there are not any consensus whether the CD4+ subsets of Tregs are as important as the CD8+ ones or not but it could be helpful to assess the CD8+ Tregs and also other Treg subtypes like Tr1 and/or Th3 in CVID patients.

Taken together, we showed that the frequency of Tregs and the expression of FOXP3 protein are reduced in CVID patients predominantly in those with autoimmune diseases. Moreover, CTLA-4 and GITR molecules are also diminished in CVID patients. Therefore, if the role of Tregs in pathogenicity of CVID disease has been verified, targeting Tregs can be considered as a therapeutic approach for CVID patients especially those with autoimmune manifestations [42]. Additionally, monitoring the Tregs' proportions and the expression of their key molecules like FOXP3 protein in conjunction with Tregs’ markers might predict that the possible autoimmune diseases may happen in future in CVID patients without autoimmunity.

Acknowledgment

This work was supported by a grant (88-04-30-9644) from Tehran University of Medical Sciences.

Ancillary