CD95 expression and function on lymphocyte subpopulations in common variable immunodeficiency (CVID); related to increased apoptosis

Abstract

Apoptosis is now recognized as a central process of development and disease, and it has been proposed as one of the mechanisms that may account for the lymphopenia seen in some diseases. In this study we measured spontaneous apoptosis and CD95 expression on different cell subpopulations from CVID patients, using flow cytometric techniques. We divided our patients into two groups according to their CD4⁺ and CD4⁺CD45RA⁺ cell counts. Our results clearly show increased spontaneous apoptosis and CD95 expression on the CD4⁺ and CD4⁺CD45RA⁺ subsets from lymphopenic CVID patients compared with normal subjects and disease controls. Interestingly, our lymphopenic CVID patients presented a profound reduction in absolute counts, mainly affecting the CD4⁺CD45RA⁺ subpopulation. We also found a statistically significant direct correlation between absolute numbers of CD4⁺CD45RA⁺ T cells and spontaneous apoptosis on the same subset in CVID patients, but attempts to induce CD95-mediated apoptosis were unsuccessful despite increased CD95 expression on CD4⁺ T cells. These findings suggest that apoptosis could be one of the mechanisms implicated in the significant lymphopenia present in these patients.

INTRODUCTION

CVID is a heterogeneous disorder characterized by hypogammaglobulinaemia, recurrent infections and impaired antibody production. Some patients demonstrate a primary defect in B lymphocyte function [1] while others show clear T cell abnormalities [2], and this is consistent with the assumption that, at least partially, the humoral deficiency is secondary to T cell defects present in these patients. A significant number of CVID patients show a profound lymphopenia. Cells chiefly contributing to this lymphopenia are CD19⁺ B cells and CD45RA⁺‘unprimed’ T CD4 cells [3]. One possible mechanism for the selective lymphopenia in CVID patients could be increased apoptosis due to abnormal or persistent activation. We have previously found that peripheral blood mononuclear cells (PBMC) and CD4⁺ T cells from some CVID patients undergo enhanced spontaneous and post-phytohaemagglutinin (PHA) activation apoptosis ‘in vitro’ when compared with normal controls [4].

The CD95 molecule is a member of the tumour necrosis factor (TNF)/nerve growth factor (NGF) receptor family [5] and it has been shown to play an important role in apoptosis [6, 7]. To explore the possibility that the lymphopenia involving different cell populations in these CVID patients may be mediated by apoptosis, we analysed spontaneous apoptosis in different cell subsets, the CD95 antigen expression in these subsets, as well as the ability of this receptor to signal apoptosis in the CD4⁺ T cell subpopulation from CVID patients and normal subjects.

To overcome the ever present difficulty of interpreting the CVID results as a consequence of the marked variability of this syndrome, we distinguish two groups of patients: one with < 600 CD4⁺ cells/mm³ and CD4⁺CD45RA⁺ < 160 cells/mm³, and the other with CD4⁺ and CD4⁺CD45RA⁺ cell counts never < 700 cells/mm³ and 300 cells/mm³, respectively.

Our results provide further data about the implications of the apoptosis mechanism in selective lymphopenia in CVID patients and confirm the important heterogeneity of this primary immunodeficiency.

PATIENTS AND METHODS

Patients and controls

We studied 14 CVID patients, eight women and six men, followed at Son Dureta Hospital. Ten patients with < 600 CD4⁺ cells/mm³ and the CD4⁺CD45RA⁺ subpopulation < 160 cells/mm³ were studied as a test group. Four CVID patients with normal lymphocyte counts (CD4⁺ and CD4⁺CD45RA⁺ cells), four X-linked agammaglobulinaemia and three IgG2 subclass deficiency patients served as a disease control.

Patients were studied at least twice with a minimum of a 6-month interval. As normal controls, 10 age-matched healthy volunteers were used. Recurrent pulmonary infections were the most common pathology seen in our patients and none of them had granulomatous disease. Two patients were diagnosed after suffering from meningococcal meningitis. All patients were receiving intravenous gammaglobulin and were free of infections at the moment of the study.

Flow cytometric analysis and antibodies

The following MoAbs were used in this study: CD4–PerCP, CD4–PE (anti-Leu-3a), CD8–PE (anti-Leu-2a), CD19–PE (anti-Leu-12), CD45RO–PE (UCHL-1) and Simultest control from Becton Dickinson (Mountain View, CA). CD45RA-RD1 (2H-4), CD95–FITC (UB-2), mouse anti-human IgG1–FITC, IgM isotype controls and anti-CD95 (CH-11) were provided by Coulter Immunotech (Marseille, France) and CD3 (UCHT1) from Dako (Glostrup, Denmark). Whole blood was used for the cytometric analysis. Two- or three-colour analysis was performed on a FACScan flow cytometer (Becton Dickinson) and data analysis was done using LYSIS II software (Becton Dickinson). A gate for the lymphocyte population was defined by forward and side light scatter characteristics, and FL3 (PerCP) strongly positive areas were further gated as CD4⁺ T cell populations for subsequent two-colour analysis.

Intracellular Bcl-2 and Bax expression

Staining of surface antigens was performed first. PBMC were permeabilized using FACS permeabilizing solution (Becton Dickinson). Bcl-2 expression was measured using direct staining with anti-Bcl-2 MoAb (FITC-labelled; Dako). Bax expression was measured using purified rabbit polyclonal antibodies (Santa Cruz Biotechnology Inc., Santa Cruz, CA). FITC-labelled swine anti-rabbit immunoglobulins (Dako) were used as the second antibody for Bax staining.

Lymphocyte culture

PBMC were obtained by Ficoll–Paque (Pharmacia, Uppsala, Sweden) density gradient centrifugation. Cells (2 × 10⁵/well) were cultured in complete medium comprising RPMI 1640 medium supplemented with 10% fetal calf serum (FCS; both from Irvine Scientific, Santa Ana, CA), 2 mml-glutamine, 100 U/ml penicillin and 100 μg/ml streptomycin in 96-well flat-bottomed plates (Nunc, Roskilde, Denmark). For analysis of spontaneous apoptosis, the cells were cultured alone for 72 h in complete medium. CD95-induced apoptosis was evaluated after 24 h in culture with anti-CD95 (CH-11) MoAb at a final concentration of 2 μg/ml. The Jurkat T cell line and PBMC from HIV-1-infected patients were used as positive controls for CD95 ligation.

PBMC from five normal controls, three CVID test patients and two CVID control patients were cultured for 3 days in medium alone or stimulated with anti-CD95 antibody during 24 h in culture. The number of viable cells remaining after the culture period was determined based on FSC and SSC profiles on the FACScan cytometer.

CD95-induced apoptosis was also evaluated in PBMC from six normal controls, three CVID test patients and three CVID control patients after prior activation for 4 days with anti-CD3 MoAb (1 μg/ml).

Measurement of apoptosis

Apoptotic cells were evaluated by flow cytometry using 7-aminoactinomycin D (7-AAD; Sigma, St Louis, MO) as DNA marker [8]. This technique allows easy quantification of apoptotic cells within specific populations. Briefly, staining of surface antigens was performed first. Ten microlitres of each MoAb were added to 2 × 10⁵ PBMC, incubated for 20 min at 4°C, washed and the cells resuspended in 500 μl PBS–FACS. Then 7-AAD (20 μg/ml) was added 15–20 min before analysis on the FACScan flow cytometer. The different cell subpopulations were gated and the apoptotic and live cells were displayed on a forward scatter versus FL3-fluorescence dot plot.

The 7-AAD⁻ cells represent live cells, the 7-AAD^dim+ cells represent early apoptotic cells and the 7-AAD^bright+ cells represent late apoptotic cells.

Because apoptosis is a dynamic process, results were expressed as the sum of the 7-AAD^bright+ and 7-AAD^dim+ cells. In some cases, cells in apoptosis were also identified with morphological techniques using May–Grünwald–Giemsa staining procedures.

Statistical analysis

The Kruskal–Wallis test (non-parametric anova) was used to determine the statistical analysis of CD95 flow cytometric expression, Bax/Bcl-2 ratios and apoptosis as the number of samples were small and more than two unpaired groups were studied. Correlation studies were done applying Spearman's rank correlation procedure and the Willcoxon test was used in order to study the differences in apoptosis after CD95 ligation.

RESULTS

CD95 expression on cell subsets from CVID patients

Circulating B lymphocytes (CD19⁺) were present in all CVID patients. In the CVID test group, all patients showed absolute numbers of CD19⁺ lymphocytes (range 20–160 cells/mm³) lower than normal controls (200–220 cells/mm³). In the CVID control group, 3/4 patients showed absolute B cell numbers lower than normal controls, and only one had normal values. As a rule, B cells were not present in the X-linked agammaglobulinaemia disease control patients, while all IgG2 deficiency patients showed decreased absolute numbers of CD19⁺ cells (range 30–100 cells/mm³).

We used flow cytometric analysis in order to examine comparatively CD95 expression on peripheral CD4⁺, CD4⁺CD45RA⁺ and CD4⁺CD45RO⁺ T cell subsets from CVID patients and control groups. CD95 expression on CD19⁺ and CD8⁺ was also studied.

As shown in Table 1, 37.1 ± 1% of total CD4⁺ T lymphocytes from normal donors were CD95⁺, whereas this percentage increased to 69 ± 5.7% in the CVID test group patients (P < 0.05) with increased CD95 mean fluorescence intensity (MFI) on CD4⁺ T cells in the CVID test group patients (P < 0.05).

Table 1.

CD95 expression (% and mean fluorescence intensity (MFI)) by CD4 T cells and CD4CD45RA and CD4CD45RO subsets from CVID test patients, normal and disease control groups†

† Numbers represent mean ± s.e.m.

* P < 0.05 compared with normal.

Three-colour immunofluorescence analysis showed that the majority of CD4⁺CD45RO⁺ peripheral T cells from healthy donors and CVID test group patients were CD95⁺ (72.5 ± 4.4% versus 82.8 ± 3.6%), without increased levels of CD95 MFI on the CVID CD4⁺CD45RO⁺ T cell population. In marked contrast, the percentage of double CD4 CD45RA/CD95-positive cells was low (7.5 ± 1%) in normal controls and significantly raised in CVID patients (26.8 ± 5.1%) (P < 0.05). CD95 MFI levels were also significantly enhanced in CD4⁺CD45RA⁺ T cells from the CVID test group compared with normal controls (7.4 ± 0.7 versus 18 ± 1.5; P < 0.05).

In the CD8⁺ subset, MFI levels and percentage of CD95⁺ cells appear raised in four patients, but the differences between the CVID test group and normal controls were not statistically significant as a group.

Neither CD95 perventage nor MFI levels on CD19⁺ cells presented differences between the CVID test group and normal controls (Table 2).

Table 2.

CD95 expression (% and mean fluorescence intensity (MFI)) by CD4 T cells and CD4CD45RA and CD4CD45RO subsets from CVID test patients, normal and disease control groups*

* Numbers represent mean ± s.e.m.

CVID, X-linked agammaglobulinaemia and IgG2 subclass deficiency disease control groups showed similar CD95 expression (percentage and MFI) to normal controls in all the populations studied.

Spontaneous apoptosis in different cell subsets from CVID patients

We evaluated spontaneous apoptosis in different cell subpopulations after 72 h in culture with solely complete medium using flow cytometric technique with 7-AAD as fluorescent DNA-binding agent. Apoptosis was clearly enhanced in the CD4⁺ (8.7 ± 1 versus 17.6 ± 1.7), CD4⁺CD45RA⁺ (7.4 ± 1 versus 19.3 ± 3.2) and CD8⁺ (8.3 ± 1.1 versus 22 ± 2.7) subpopulations from CVID test patients compared with the normal control groups. A representative experiment is shown in Fig. 1.

Fig. 1.

A representative experiment of flow cytometric analysis of apoptosis on CD4⁺ T and CD4⁺CD45RA⁺ T cells from normal subjects and CVID lymphopenic patients employing the 7-amino-actinomycin D (7-AAD) technique.

Normal levels of spontaneous apoptosis were observed in the CD4⁺CD45RO⁺ and CD19⁺ B lymphocytes. Apoptosis in these lymphocyte subpopulations was not increased when CVID, X-linked agammaglobulinaemia and IgG2 subclass deficiency disease control groups were studied (Fig. 2).

Fig. 2.

Percentage of spontaneous apoptosis within CD4⁺ and CD8⁺ (a) T cells and CD4 CD45RA⁺ and CD4 CD45RO⁺ (b) T cell subsets after 72 h on culture. Each symbol represents a different subject. The horizontal lines represent mean of apoptotic cells for each group. The statistically significant differences were found between normal controls and CVID test patients on CD4⁺, CD8⁺ and CD4 CD45RA⁺ T cells. (P < 0.05, Kruskal–Wallis test.)

When cell recovery was investigated, we found a reduction in recovery of viable cells in the CVID test group after 72 h in culture in medium alone (recovery 60 ± 4%) compared with normal controls (recovery 85 ± 5%) (Table 3). Thus, we found increased spontaneous apoptosis in some CVID cell subpopulations, and on the other hand we also found reduced cell numbers after 72 h in culture. These data suggest that the observed apoptosis did result in cell loss. However, neither a reduction in recovery of viable cells after stimulation with anti-CD95 during 24 h in culture nor increased CD95-induced apoptosis were found.

Table 3.

Kinetics of cell numbers in cultured peripheral blood mononuclear cells (PBMC) from normal controls and CVID patients*

* Numbers represent mean ± s.e.m. Cell number (× 10⁶).

d.c., Disease control.

We also quantified Bax/Bcl-2 ratios in CD4⁺ T cells both before and after 72 h in culture to determine if any changes occurred. The differences between the CVID test group and normal controls were not statistically significant (Table 4).

Table 4.

Mean fluorescence intensities (MFI) for Bcl-2 and Bax expression and Bax/Bcl-2 ratios in CD4 T cells from CVID test patients, normal and disease control group, before and after 72 h in culture*

* Numbers represent mean ± s.e.m.

d.c., Disease control.

Correlation between CD95 antigen expression, apoptosis and absolute numbers of T cell subsets

We also analysed the correlation between CD95 antigen expression (% and MFI), absolute numbers and spontaneous apoptosis on different subsets from CVID patients. Statistically significant correlations were found in the CD4⁺CD45RA⁺ subset and not in the other subpopulations. We also found a statistically significant inverse correlation between the absolute numbers of the CD4⁺CD45RA⁺ subset in peripheral blood and CD95 antigen expression (percentage and MFI) of the same subset. The inverse correlation between spontaneous apoptosis of the CD4⁺CD45RA⁺ subset and its absolute numbers was also statistically significant. We also found a statistically significant direct correlation between spontaneous apoptosis of the CD4⁺CD45RA⁺ subpopulation and CD95 antigen expression (percentage and MFI) of the same subset (Fig. 3).

Fig. 3.

Correlations between percentage of CD95 antigen-positive CD4 CD45RA⁺ T cells and its absolute numbers (a), percentage of CD95 antigen-positive CD4 CD45RA⁺ T cells and spontaneous apoptosis on the same subset (b). Each r and P value is shown (Spearman's rank correlation test).

CD95-induced apoptosis

To study the induction of apoptosis after CD95 ligation, we incubated PBMC from CVID patients (test and disease control), X-linked agammaglobulinaemia, IgG2 subclass deficiency and normal controls, with anti-CD95 (CH-11 MoAb) at a final concentration of 2 μg/ml during 24 h in culture. In the same experiment, we cultured Jurkat T cell line and PBMC from HIV⁺ patients as positive controls. As we expected, significant levels of apoptosis were induced in CD4⁺ T cells from HIV⁺ patients and Jurkat T cell line cells. However, despite over-expression of CD95 on CD4⁺ T CVID test patients' cells, we were unable to induce apoptosis using this protocol of CD95 ligation. This failure to induce apoptosis was also observed in CVID disease controls as well as for X-linked agammaglobulinaemia, IgG2 subclass deficiency and normal controls (Table 5).

Table 5.

Percentage of CD95-induced apoptosis on CD4 T cells from CVID test patients and disease control groups†

† Numbers represent mean ± s.e.m.

* P < 0.05 compared with spontaneous apoptosis.

d.c., Disease control.

We also activated the cells from CVID patients with anti-CD3 antibody for 4 days and then stimulated the cells with anti-CD95 during 24 h. Although we studied a small number of patients, CVID test patients clearly did not show an increase in CD95-induced apoptosis using this protocol of activation, in contrast with CVID disease and normal controls (Table 6).

Table 6.

Percentage of CD95-induced apoptosis on CD4 T cells from CVID test patients and disease control group prior to activation with anti-CD3 for 4 days

All values are shown as mean ± s.e.m.

d.c., Disease control.

* The cells from CVID patients and normal controls were activated with anti-CD3 antibody (1 μg/ml) for 4 days.

† The cells from CVID patients and normal controls were activated with anti-CD3 antibody (1 μg/ml) for 4 days and then stimulated with anti-CD95 (2 μg/ml) for 24 h.

DISCUSSION

In CVID patients, besides a failure in B cell differentiation an array of T cell abnormalities has been described. Many CVID patients show a profound lymphopenia in B and CD4⁺ T cell populations, mainly in the CD4⁺CD45RA⁺ subset.

The aim of this study was to investigate CD95 expression and the degree of spontaneous apoptosis in vitro in different peripheral lymphocyte subpopulations from CVID patients. Because CVID is very much a heterogeneous syndrome, we arbitrarily divided CVID patients into two groups in accordance with their CD4⁺ and CD4⁺CD45RA⁺ T cell absolute counts.

Our findings reveal that the great majority of CD4⁺ T cells and CD4⁺CD45RA⁺ subset lymphocytes from lymphopenic CVID patients express the CD95 antigen, compared with normal and disease control groups. Furthermore, we performed quantitative studies on relative CD95 antigen densities, measured as MFI levels, and our data clearly indicate that CD95 antigen MFI is significantly enhanced in CD4⁺ T cells and the CD4⁺CD45RA⁺ subpopulation. No alterations in the percentage of CD4⁺ CD45RO⁺, CD19⁺ and CD8⁺ cells expressing CD95, or in MFI antigen levels, were observed. Thus, our data demonstrate that CD95 antigen (MFI and percentage) on CD4⁺ T cells and the CD4⁺CD45RA⁺ subset, is enhanced in a group of lymphopenic CVID patients compared with healthy controls, CVID patients without lymphopenia, patients with X-linked agammaglobulinaemia and IgG2 subclass deficiency, suggesting that enhanced CD95 expression could have an important role in CD4⁺CD45RA⁺ lymphopenia seen in these CVID patients.

In addition, we describe an increased spontaneous apoptosis of CD4⁺ T cells and the CD4⁺CD45RA⁺ subset as well as of CD8⁺ T cells from lymphopenic CVID patients compared with control groups.

Our results are partially in agreement with the work of Saxon et al. [9], who found increased CD95 antigen expression and spontaneous apoptosis in both B and T cells from a selected subpopulation of CVID patients whose B cells demonstrated a CD20^bright/l-selectin (CD62L)^dim phenotype. Our results show increased CD95 expression only on CD4⁺ cells (mainly on CD4⁺CD45RA⁺ cells) but not on CD19⁺ cells. Perhaps the discrepancy is based on the different groups of patients we studied, since our patients did not show the characteristic phenotype of Saxon's patients (data not shown).

The CD95 antigen is a protein expressed on the surface of lymphocytes which belong to NGF/TNF receptor superfamily. Signalling through CD95 has been demonstrated to induce apoptosis of transformed cell lines and chronically activated T cell clones [10–12]. Expression of CD95 on activated peripheral T and B cells may account for their susceptibility to apoptosis, although freshly isolated CD95⁺ T cells in particular are insensitive to several apoptosis-inducing agents [13, 14].

The increased susceptibility to apoptosis and enhanced CD95 expression on different subpopulations have been suggested as events that could contribute to several pathologies, e.g. systemic lupus erythematosus (SLE) [15–17] and HIV infection [18–20], although the molecular mechanisms underlying these processes are unknown.

Although the aetiopathogeny of HIV infection is well known, also that important differences exist between this pathology and CVID, there are many similarities between them. Both pathologies presents CD4⁺ T and CD4⁺CD45RA⁺ lymphopenia [1, 3], functional abnormalities of T lymphocytes [21, 22] and continuous immune activation in vivo [23, 24]. Increased CD95 expression and enhanced apoptosis have been described in HIV infection, where different authors have demonstrated increased CD95 expression (% and/or MFI) on CD4⁺ T cells and the CD4⁺CD45RA⁺ subpopulation [20, 25, 26]. In HIV-infected patients several authors have found that the increased susceptibility of T cells to undergo apoptosis may involve a CD95 pathway, and an inverse correlation has been described between peripheral blood CD4⁺ T cell counts with spontaneous and with CD95-induced apoptosis on the same subpopulation, indicating that apoptosis may be contributing to the characteristic CD4⁺ T depletion seen in HIV infection [19, 20].

The fact that we had found an inverse correlation between apoptosis levels and CD95 expression, and CD95 expression and the absolute counts of the CD4⁺CD45RA⁺ subpopulation (but not in other subsets), together with a direct correlation between absolute numbers and apoptosis for the same subpopulation, would give support to the hypothesis that CD95-mediated apoptosis may account for the important CD4⁺CD45RA⁺ lymphopenia present in these lymphopenic CVID patients.

T cells from CVID patients show a persistent immune activation in vivo, and for this reason we tried to induce CD95-mediated apoptosis without prior activation with mitogenic stimuli. We were not able to induce CD95-mediated apoptosis from CVID lymphopenic patients in spite of increased CD95 expression on CD4⁺ T cells, but when we activated the cells from CVID patients with anti-CD3 antibody before ligating CD95 we were again not able to induce CD95-mediated apoptosis.

The impossibility of inducing CD95-mediated apoptosis suggests several interpretations. CD4⁺ T cells from lymphopenic CVID patients may have an unfunctional CD95 antigen unable to activate the death process, although Saxon et al. [9] have described enhanced CD95-mediated apoptosis in Epstein–Barr virus (EBV)-transformed cell lines from CVID patients, suggesting that CD95 antigen may drive the apoptotic signal, at least in CD19⁺ cells.

Other molecules besides CD95 have been described to be involved in the apoptosis pathway, especially Bcl-2 gene family members [27–29]. We have not found differences in Bax/Bcl-2 ratios in CD4⁺ T cells between CVID lymphopenic patients and normal controls.

Another possibility is that CVID patients could have increased serum levels of soluble CD95 resulting in neutralization of the antibody to CD95 and, consequently, in the inhibition of CD95-mediated apoptosis. Although we did not carry out these studies in our CVID patients, this has been described in SLE patients who also show no cell death after CD95 stimulation [30, 31]. Altered cytokine profiles may affect immune responses. In HIV infection, type 1 cytokines were shown to have a protective role in preventing apoptosis, whereas type 2 cytokines were not protective. CD95-mediated apoptosis of CD4⁺ T cells was prevented by the addition of IL-12 in HIV patients [32]. It has been described by North et al. [33] that there are high levels of cells able to make interferon-gamma (IFN-γ) in CVID CD4⁺ T cells, and it is interesting to note that we have described for the first time increased IL-12 serum levels in CVID patients [34].

Perhaps the increased CD95 expression that we found in the CD4⁺CD45RA⁺ subset from lymphopenic CVID patients was the result of chronic immune activation that has been described by several authors in CVID patients [23, 24, 35], and that other pathways independent of CD95 may be involved in the apoptotic process, for instance, the TNF system. Interestingly, Aukrust et al. [36] have found glutathione abnormalities both in plasma and in CD4⁺CD45RA⁺ T lymphocytes and monocytes, but not in CD19⁺ B lymphocytes from lymphopenic CVID patients, and these glutathione redox disturbances were correlated with activation of the TNF system in CVID. On the other hand, increased oxidative stress has been related to enhanced apoptosis [37, 38].

In conclusion, our study shows that T cells from lymphopenic CVID patients undergo enhanced spontaneous apoptosis, with the CD4⁺CD45RA⁺ T cells having the most severe abnormalities: decreased absolute numbers, increased CD95 expression and apoptosis levels. Perhaps an inappropriate CD4⁺CD45RA⁺ T cell activation or costimulatory signals not adequately provided by accessory cells trigger the apoptotic process, apparently independently of CD95.