Characterizing a soluble survival signal for activated lymphocytes from CD14⁺ cells

Abstract

T-cell activation requires at least two signals: antigen and a costimulatory signal. As antigen-presenting cells play an important role in this area, the role of CD14⁺ cells in T-cell activation, proliferation and activation-induced cell death (AICD) was investigated. Using phytohaemagglutinin (PHA) activation, it was found that CD14⁺ cell depletion resulted in significantly greater AICD, decreased lymphocyte growth and up-regulated interleukin-2 (IL-2) secretion. However, T-cell activation was delayed according to the expression of CD69 and CD25. Dynabeads conjugated with anti-CD14 monoclonal antibody (mAb) bound CD14⁺ cells and induced secretion of IL-1β, tumour necrosis factor-α (TNF-α), transforming growth factor-β (TGF-β) and IL-6, but not IL-2, IL-12 or IL-15. Supernatants were collected from Dynabeads-activated CD14⁺ cell cultures and designated as ‘CD14 cocktails’. Addition of CD14 cocktails to CD14⁺ cell-depleted mononuclear cell cultures reversed the increased AICD, decreased lymphocyte growth and increased IL-2 secretion. Depletion of IL-1β and TNF-α in the CD14 cocktails by panning followed by blocking with the corresponding mAbs had no effect on the active AICD protection. TGF-β was determined not to be the active factor owing to the presence of >1·0 ng of TGF-β in the media for culturing both CD14⁺ and CD14⁻ peripheral blood mononuclear cells (PBMC). The CD14 cocktails did not contain IL-12 and IL-15. Depletion of IL-6 with panning followed by blocking residual IL-6 with anti-IL-6 mAb significantly reduced the protective effect of the CD14 cocktails. Human recombinant IL-6 also partially reversed the effects of CD14⁺ cell depletion on AICD, lymphocyte growth and IL-2 secretion. The data suggest that IL-6 is one of the active factors in the survival signal from CD14⁺ cells.

Introduction

Currently, two mechanisms have been reported to contribute to peripheral immune tolerance: clonal deletion and clonal anergy. Peripheral clonal deletion occurs after mature T cells encounter antigen and become activated. Two events may contribute to peripheral clonal deletion: T-cell removal by activation-induced cell death (AICD or active apoptosis), or passive apoptosis that occurs after removal of stimulating antigens and subsequent removal of growth factors.¹ AICD has been considered an important mechanism for maintaining peripheral tolerance.^2–6 However, it may play a role in autoimmune disease, where autoreactive lymphocytes may be more resistant to apoptosis.^7–9 Recently, intraperitoneal injection of soluble myelin proteins was reported to induce apoptosis of autoreactive T cells in an experimental allergic encephalomyelitis (EAE) animal model.¹⁰ It is possible that specific mechanisms exist to keep activated lymphocytes alive for longer by avoiding AICD, thus maintaining the immune response. However, such mechanisms may promote autoreactivity in genetically predisposed individuals by causing accumulation and expansion of autoreactive lymphocytes. Several factors and cytokines have been reported to facilitate lymphocyte apoptosis: interferon-γ (IFN-γ),^11,12 tumour necrosis factor-α (TNF-α),¹³ interleukin (IL)-2¹⁴ and soluble (s)CD137.¹⁵ Alternatively, several other factors and cytokines have been reported to prevent lymphocyte apoptosis: IL-2,¹⁶ TNF-α,¹⁷ IL-1,¹⁸ IL-6,¹⁹ IL-7,²⁰β-chemokines²¹ transforming growth factor-β (TGF-β),^22,23 IL-15²⁴ and IFN-α/β.^25–27 There are contrasting data on the role of monocytes in lymphocyte apoptosis. While monocytes were reported to rescue activated lymphocytes from spontaneous apoptosis,^28,29 other laboratories reported that monocytes were required to prime T cells to undergo apoptosis.³⁰ Soluble factors from activated monocytes were reported to have no effect on apoptosis.^28,29

Professional antigen-presenting cells (APCs) can trigger immune responses. APCs include monocytes, macrophages, dendritic cells and B cells. A common feature of these APCs is that they are CD14 positive, at least in the naïve state.^31–33 The CD14 molecule has important functions in innate immunity, including binding bacteria, lipopolysaccharide (LPS) and virus, and also in the clearance of apoptotic bodies.^34,35 Therefore, CD14⁺ cells play a critical role in innate and acquired immunity by binding antigens, presenting antigens to T cells and removing apoptotic cells after immune response. T-cell activation requires at least two signals.^36,37 Upon activation, T cells will leave local lymph nodes where the activation occurred and migrate to the sites where antigens broke the first barrier of the immune system and invaded tissues. The activated lymphocytes have to stay alive during their migration. The life and death of activated lymphocytes is therefore an important checkpoint controlling an immune response. The survival of activated lymphocytes is required for T lymphocytes to have sufficient time to perform an immune response and to generate immune memory. However, in genetically programmed individuals, the survival of autoreactive lymphocytes could cause autoimmune disease. The death of activated lymphocytes is equally important in ending an immune response after antigens have been removed and in removing autoreactive lymphocytes. However, early death of activated lymphocytes could cause an aborted immune response. The detailed mechanisms for controlling the life and death of activated lymphocytes are still unknown.^38,39 Herein, we postulate that APCs are able to provide essential signals to T lymphocytes, which maintain an effective immune response by preventing AICD.

In this study, phytohaemagglutinin-p (PHA-p) was used to activate lymphocytes in the presence or absence of CD14⁺ cells. Soluble factors from Dynabeads-activated CD14⁺ cells were studied to analyse their effects on AICD. The findings suggested that IL-6 is one of the active factors in the CD14 cocktails that prevented apoptosis in order to maintain an active immune response.

Materials and Methods

Study population

All subjects (n = 40) were female or male (age-range 35–60 years). No subject enrolled in this study had any chronic or acute diseases. Peripheral blood was drawn into heparinized tubes for all analyses.

Isolation of peripheral blood mononuclear cells from normal control subjects and stimulation with PHA-p

Peripheral blood mononuclear cells (PBMC, or CD14⁺ PBMC) were isolated from whole blood using Histopaque.⁴⁰ After isolation, mononuclear cells were reconstituted with RPMI-1640 [containing 2·83 mg/ml HEPES, 2 mg/ml NaHCO₃, 100 U/ml penicillin, 100 µg/ml streptomycin, 3·125×10⁻⁵ 2-mercaptoethanol (2-ME) and 10% fetal bovine serum (FBS)], counted and aliquoted for different experiments. The aliquoted cells were then adjusted to 1×10⁶ cells/ml with RPMI-1640 containing 10 µg/ml PHA-p (Sigma, St Louis, MO), seeded into six- or 24-well culture plates (Costar, High Wycombe, UK), depending on the experiments, and cultured at 37° in 5% CO₂.

CD14⁺ cell depletion using Dynabeads and preparation of CD14 cocktails

CD14⁺ cells were removed from original PBMC (CD14⁺ PBMC) by Dynabeads (M-450 CD14; Dynal Inc., Oslo, Norway) according to the manufacturer's recommendation. Cells were separated into CD14⁻ PBMC and Dynabeads-bound cells (CD14⁺ cells) using a magnetic particle concentrator (MPC-1; Dynal). The CD14⁻ PBMCs were then tested for CD14 antigen compared with CD14⁺ PBMC.

The Dynabeads-bound cells (CD14⁺ cells) were washed three times with phosphate-buffered saline (PBS), pH 7·4, and adjusted to 1×10⁶ cells/ml using macrophage serum-free medium (Gibco BRL, Carlsbad, CA). The cells were cultured at 37° in 5% CO₂ and the supernatants collected after 18 hr. The supernatants were designated as ‘CD14 cocktails’.

Flow cytometric analysis of apoptosis and cell-surface antigens

Cells were adjusted to 4×10⁶/ml in PBS (pH 7·4 with 5% FBS). Cells (65 µl) were added to 5-ml tubes and directly fluorescenated antibodies were added into the tubes. The tubes were incubated for 30 min at 4°, washed once with PBS (pH 7·4) and reconstituted in PBS (pH 7·4). The cells were analysed using flow cytometry (three colors). Antibodies used included: CD3-phycoerythrin-cyanin 5.1 (PC5), T4-fluorescein isothiocyanate (FITC), CD4-phycoerythrin, red dye 1 (RD1), T8-FITC, CD8-RD1, B4-RD1 (CD19), B4-FITC (CD19), CD69-phycoerythrin (PE), CD25-PC5, CD122-PE, CD132-PE, CD95-PE and CD14-PE (purchased from Beckman Coulter, Miami, FL). For apoptosis, the Annexin V-FITC kit was used (Immunotech, Beckman Coulter) to stain the cells. Briefly, 4×10⁵ cells were washed once with PBS and adjusted to a concentration of 4×10⁶/ml with 100 µl of buffer; 1 µl of Annexin V and 5 µl PI (propidium iodide) were added and mixed. The cells were then incubated at 4° for 10 min in the dark and analysed by flow cytometry without washing.

CD14 supernatant studies

For the AICD protection assay, CD14⁻ PBMC were adjusted to 1×10⁶ cells/ml with complete RPMI-1640 containing PHA (10 µg/ml) and either 10%, 20%, 40%, 80% or 100% CD14 cocktails. CD14⁺ PBMC were used for comparison. The cells were seeded into 96- or 24-well culture plates, depending on the experiment, and cultured at 37° in 5% CO₂. After 24, 48 and 72 hr, cells were collected and analysed for apoptosis by flow cytometry.

IL-1β, TNF-α and IL-6 depletion assays

IL-1β (catalogue no. DLB50; R & D Systems, Minneapolis, MN), TNF-α (catalogue no. DTA50; R & D Systems) and IL-6 enzyme-linked immunosorbent assay (ELISA) kits (catalogue no. D6050; R & D Systems) were used. Briefly, CD14 cocktails were added to the ELISA plates at 100 µl/well and incubated at 4° for 2 hr. After the first 2-hr incubation, the CD14 cocktails were transferred into new wells for incubation at 4° for another 2 hr. After the second 2-hr incubation, the CD14 cocktails were transferred into new wells again and were subjected to a third round of depletion (2 hr).

IL-1β, TNF-α and IL-6 blocking assays

For the IL-1β blocking assay, anti-IL-1β monoclonal antibody (mAb) [(R & D Systems) ND₅₀ (50% neutralizing dose) and ND₁₀₀ (100% neutralizing dose) are ≈ 0·001–0·003 µg/ml and 0·1–1 µg/ml, respectively, when the IL-1β concentration is 50 pg/ml] was used to block the residual IL-1β. Briefly, the IL-1β-depleted CD14 cocktails were further incubated with 9·6 µg/ml anti-IL-1β antibody at 37° in 5% CO₂ for 1 hr before it was used for the protection assay. For the TNF-α blocking assay, anti-TNF-α mAb (R & D Systems, ND₅₀ and ND₁₀₀ are ≈ 0·02–0·04 µg/ml and 0·1–1 µg/ml, respectively, when the TNF-α concentration is 250 pg/ml) was used to block the residual TNF-α. Briefly, the TNF-α-depleted CD14 cocktails were further incubated with anti-TNF-α antibody (6·4 µg/ml) at 37° in 5% CO₂ for 1 hr before it was used for the protection assay. For the IL-6-depletion assay, anti-IL-6 mAb (R & D Systems, ND₅₀ and ND₁₀₀ are ≈ 0·05–0·15 µg/ml and 1·0–5 µg/ml, respectively, when the IL-6 concentration is 2·5 ng/ml) was used to block the residual IL-6. Briefly, the IL-6-depleted CD14 cocktails were further incubated with anti-IL-6 antibody (10 µg/ml) at 37° in 5% CO₂ for 1 hr before it was used for the protection assay.

ELISA detection of cytokine secretion

IL-1β (catalogue no. DLB50, R & D Systems; the minimum detectable dose was typically less than 1 pg/ml), TNF-α (catalogue no. DTA50, R & D Systems; the minimum detectable dose was typically less than 4·4 pg/ml), TGF-β (catalogue no. DB100, R & D Systems; the minimum detectable dose was typically less than 7 pg/ml), IL-6 (catalogue no. D6050, R & D Systems; the minimum detectable dose was typically less than 0·70 pg/ml), IL-2 (catalogue no. D2050, R & D Systems; the minimum detectable dose was typically less than 7 pg/ml), IL-12 (catalogue no. D1200, R & D Systems; the minimum detectable dose was typically less than 5·0 pg/ml) and IL-15 (catalogue no. D1500, R & D Systems; the minimum detectable dose was typically less than 2 pg/ml) were used to test cytokine contents in the CD14 cocktails and in the supernatants of the PHA-activated lymphocytes.

Statistical analysis

SPSS software (statistical package for the social sciences, version 10.1 for windows) was used to perform the statistical analyses. The statistical methods used in this study included the paired-sample t-test, independent-sample t-test and analysis of variance (anova).

Results

Depletion of CD14⁺ cells from PBMC resulted in significantly higher activation-induced cell death

To demonstrate the role of CD14⁺ cells in lymphocyte function, CD14⁺ cells were depleted by Dynabeads (M-450, CD14) that were conjugated with anti-CD14 mAbs. CD14⁺ cells were virtually undetectable after two rounds of depletion (data not shown). However, Dynabeads did not deplete B cells (data not shown). The effects of CD14⁺ cell depletion on AICD were then evaluated by stimulating CD14⁺ and CD14⁻ PBMC with PHA-p (10 µg/ml). Figure 1(a) shows that depletion of CD14⁺ cells resulted in significantly higher AICD rates at 24, 48 and 72 hr. To determine if the observed PHA-p-induced AICD is a typical process of apoptosis, first, activated lymphocytes were stained with both PI and annexin V (Fig. 1b, upper and lower panels show CD14⁺ PBMC and CD14⁻ PBMC that were stimulated with PHA-p for 0, 24, 48 and 72 hr, respectively). The data suggest that PHA-p stimulation caused a low level of AICD in the presence of CD14⁺ cells (upper panel), the depletion of CD14⁺ cells resulted in increased AICD (lower panel). It also shows that most of the dead cells were PI/annexin V double-positive cells, suggesting that the AICD is a process of apoptosis. To further prove that AICD is typical of the apoptotic process, DNA was extracted from PHA-p-activated CD14⁻ PBMC at 0, 24, 48 and 72 hr. DNA ladders were visualized by electrophoresis. As shown in the upper panel of Fig. 1(c), CD14⁻ PBMC demonstrates a DNA ladder at 48 and 72 hr (lanes 3 and 4). No DNA ladders occurred at 0 and 24 hr (lanes 1 and 2). DNA was therefore extracted from both CD14⁺ and CD14⁻ PBMC at 72 hr. As shown in the lower panel of Fig. 1(c), both PHA-p-stimulated CD14⁺ and CD14⁻ PBMC demonstrated a DNA ladder at 72 hr (lanes 3 and 5). No DNA ladders occurred at 0 hr in either CD14⁺ or CD14⁻ PBMC (lanes 2 and 4). Lane 6 shows positive controls (apoptotic U937 cells). CD14⁺ PBMC and CD14⁻ PBMC were also analysed by electron microscopy. As shown in the upper panel of Fig. 1(d), typical apoptotic nuclei did not occur in PHA-p-activated CD14⁺ PBMC at 24 hr. However, typical apoptotic nuclei were seen in CD14⁻ PBMC stimulated with PHA-p for 24 hr (lower panel of Fig. 1d). The nuclei were shrunken. The heterochromatins were condensed and usually coalesced against one pole of the nuclear membrane (arrow in the lower panel of Fig. 1d). The above data proved that AICD is a process of apoptosis.

Figure 1.

Depletion of CD14⁺ cells resulted in increased activation-induced cell death (AICD) rates. CD14⁺ peripheral blood mononuclear cells (PBMCs) were isolated by Histopaque and depleted of CD14⁺ cells by Dynabeads (CD14⁻ PBMC). CD14⁺ and CD14⁻ PBMC were stimulated with phytohaemagglutinin-p (PHA-p) (10 µg/ml) at a cell concentration of 1×10⁶/ml. (a) At 24, 48 and 72 hr, CD14⁺ and CD14⁻ PBMC were analysed by flow cytometry. Propidium iodide (PI)-positive cells were used to quantify the apoptotic rate. *P < 0·0001: CD14⁺ versus CD14⁻, n = 18; paired sample t-test. Bars show the mean value±1 SE. (b) At 24, 48 and 72 hr, CD14⁺ and CD14⁻ PBMC were analysed by flow cytometry using annexin V-fluorescein isothiocyanate (FITC) and PI staining. (c) Upper panel: kinetic pattern of the DNA ladder in CD14⁻ PBMC stimulated with PHA-p. M, Marker; lane 1, 0 hr; lane 2, 24 hr; lane 3, 48 hr; lane 4, 72 hr. Lower panel: comparison of the DNA ladder in CD14⁺ and CD14⁻ PBMC activated with PHA for 72 hr. Lane 1, marker; lane 2, CD14⁺ PBMC at 0 hr (negative control); lane 3, CD14⁺ PBMC at 72 hr; lane 4, CD14⁻ PBMC at 0 hr (negative control); lane 5, CD14⁻ PBMC at 72 hr; lane 6, positive control (apoptotic U937 cells). (d) CD14⁺ (upper panel) and CD14⁻ PBMC (lower panel) stimulated with PHA-p for 24 hr were washed three times in 0·1 m sodium cacodylate buffer, pH 7·2, fixed in 3% glutaraldehyde for 2 hr and stored in 0·1 m sodium cacodylate buffer, pH 7·2. The fixed cells were analysed by electron microscopy. (Panels (c) and (d) on following page)

The effects of CD14⁺ cell depletion on lymphocyte activation and proliferation

We next examined whether CD14⁺ cell depletion affected lymphocyte activation. As shown in Fig. 2(a), CD69 (early activation antigen) expression in CD14⁻ PBMC is decreased at 24 hr and increased at 72 hr as compared with CD14⁺ PBMC. CD25, CD122 and CD132 (IL-2 receptor α, β and γ chain, respectively) were also used to evaluate the lymphocyte activation. Figure 2(b) shows that CD25 expression is decreased at 24 and 48 hr but recovers by 72 hr. The effect of CD14⁺ cell depletion on CD122 and CD132 expression is not significant (P > 0·05). The above data suggest a delay of T-cell activation in the absence of CD14⁺ cells.

Figure 2.

Effects of CD14⁺ cell depletion on lymphocyte activation and proliferation. CD14⁺ peripheral blood mononuclear cells (PBMC) were isolated by Histopaque and depleted of CD14⁺ cells by Dynabeads (CD14⁻ PBMC). CD14⁺ and CD14⁻ PBMC were then stimulated with phytohaemagglutinin-p (PHA-p) (10 µg/ml) at a concentration of 1·0×10⁶ cells/ml (a) CD14⁺ and CD14⁻ PBMC were analysed by flow cytometry at 24, 48 and 72 hr for the expression of CD69 antigen. *P < 0·01: CD69/CD14⁻ PBMC versus CD69/CD14⁺ PBMC at 24 and 72 hr. †P < 0·05: CD3/69/CD14⁻ PBMC versus CD3/69/CD14⁺ PBMC at 72 hr. n = 10 (CD69); n = 5 (CD3/69); independent sample t-test. The differences of CD69 expression at 48 hr and the differences of CD3/69 expression at 24/48 hr were not significant (P > 0·05). Bars show the mean value±1 SE. (b) CD14⁺ and CD14⁻ PBMC were analysed by flow cytometry for the expression of CD25, CD122 and CD132 antigens. *P < 0·01: CD14⁺ (24 hr) versus CD14⁻(24 hr); paired t-test, n = 8. CD25, CD122 and CD132 expression before and after CD14⁺ cell depletion at all the other time-points were not significant. Bars show the mean value±1 SE. (c) Supernatants from CD14⁺ and CD14⁻ PBMC cultures were collected, aliquoted and stored at −80°. Enzyme-linked immunosorbent assay (ELISA) was used to test for interleukin-2 (IL-2) production. *P < 0·05; †P < 0·01: CD14⁺ versus CD14⁻; n = 8; paired sample t-test. Bars show the mean value±1 SE. (d) At 24, 48 and 72 hr, viable cells were counted as Trypan blue-excluded cells. Cell death was analysed by flow cytometry using annexin V-fluorescein isothiocyanate (FITC) and -propidium iodide (PI) staining, and by Trypan blue staining. The data represent the mean of three independent experiments. Hatched bars represent cell death rates and lines represent lymphocyte counts. CD14⁺/PI & Annexin and CD14⁻/PI & Annexin: cell death in CD14⁺ and CD14⁻ PBMC stained with PI/Annexin and analysed by flow cytometry; CD14⁺/TB and CD14⁻/TB: cell death in CD14⁺ and CD14⁻ PBMC stained with Trypan blue and analysed by light microscopy; CD14⁺/count and CD14⁻/count: viable lymphocyte count stained with viable excluding dye (Trypan blue) and analysed by light microscopy.

The IL-2 secretion before and after CD14⁺ cell depletion was also examined. Supernatants from PHA-activated CD14⁺ PBMC and CD14⁻ PBMC were collected at 24, 48 and 72 hr and tested for IL-2 production. As shown in Fig. 2(c), supernatants from cultures of the PHA-p-activated CD14⁻ PBMC contained significantly higher levels of IL-2 compared with that from cultures of the PHA-activated CD14⁺ PBMC.

Lymphocyte growth was evaluated by cell counts using Trypan blue (a viable excluding dye) in each well. Cell death was quantified by PI/annexin V and Trypan blue staining. Figure 2(d) shows the relationship between lymphocyte growth and cell death in each well in the presence (CD14⁺) or absence (CD14⁻) of CD14⁺ cells. Each experiment started with a total of 1×10⁶ lymphocyte/well. In both situations, lymphocyte counts decreased significantly during the first 24 hr, started to increase at 24 hr and showed a dramatic increase between 48 and 72 hr. However, in the absence of CD14⁺ cells (open triangle) each well had fewer lymphocytes at each time-point as compared with in the presence of CD14⁺ cells (closed triangle). Consistent with the lymphocyte growth, the data suggested that AICD peaked at 24 or 48 hr in both situations according to PI/annexin V and Trypan blue staining (hatched bars). Cell death rates in the cultures were decreased at 72 hr. However, in the absence of CD14⁺ cells the AICD rates were much higher at each time-point as compared with in the presence of CD14⁺ cells. The percentages of lymphocyte subsets (CD3⁺, CD4⁺, CD3/4⁺ and CD3/8⁺) changed in a similar pattern during the 3-day cultures in the presence or absence of CD14⁺ cells, as shown by flow cytometric analysis (data not shown).

CD14 cocktails reversed the effects of CD14⁺ cell depletion on AICD, lymphocyte growth and IL-2 secretion

To examine the potential role of ‘soluble factors’ in AICD, CD14 cocktails were prepared as described in the Materials and methods. The cytokine content was tested using sandwich ELISA. As shown in Fig. 3(a), the CD14 cocktails contained IL-1β, TNF-α, TGF-β and IL-6. No IL-2, IL-12 or IL-15 was found, implicating the activation of monocytes and secretion of monocytic cytokines. Cytokine concentrations peaked at ≈ 18 hr (data not shown). Therefore, the CD14 cocktails were collected at 17–18 hr and stored at −80° for future studies.

Figure 3.

CD14 cocktails protect activated lymphocytes from undergoing activation-induced cell death (AICD). Dynabeads were mixed with CD14⁺ peripheral blood mononuclear cells (PBMC) at a ratio of 50 µl/10⁷ cells and incubated at 4° for 1 hr with gentle rotation. After 1 hr, free Dynabeads and Dynabeads-bound cells were collected using a magnetic particle concentrator (Dynal, MPC). The Dynabeads and Dynabeads-bound cells were resuspended in serum-free medium. (a) CD14 cocktails were collected at 18 hr after incubation of Dynabeads-bound cells (CD14⁺ cells) with Dynabeads. The interleukin (IL)-1β, tumour necrosis factor-α (TNF-α), transforming growth factor-β (TGF-β), IL-6, IL-2, IL-12 and IL-15 enzyme-linked immunosorbent assay (ELISA) kits were used to test the cytokine contents. Bars show the mean value±1 SE. (b) CD14⁺ PBMC, CD14⁻ PBMC and CD14⁻ PBMC with addition of different doses of CD14 cocktails were cultured at 37° in 5% CO₂. At 0, 24 and 48 hr time-points, cells were analysed by flow cytometry and the apoptotic rates were quantified as propidium iodide (PI)-positive cells. The data represents the mean of three independent experiments. Bars show the mean value±1 SE. (c) CD14⁺ PBMC, CD14⁻ PBMC and CD14⁻ PBMC with addition of 20% CD14 cocktails were cultured at 37° in 5% CO₂. At 24, 48 and 72 hr time-points, apoptosis was analysed as described in (b). *CD14⁺ versus CD14⁻: P < 0·001 at all time-points; ‡CD14⁻ versus CD14⁻/CD14CK: P < 0·001 at all time-points; AICD rates between CD14⁺ and CD14⁻/CD14CK were not significant at any time-point. n = 12, analysis of variance (anova). Bars show the mean value ± 1 SE. (d) CD14⁺ and CD14⁻ PBMC were stimulated with phytohaemagglutinin-p (PHA-p) at a concentration of 1×10⁶ cells/ml. CD14 cocktails (20%, vol/vol) were added to some of the CD14⁻ PBMC at time 0. At 24, 48 and 72 hr time-points, cell numbers were counted in each well. Viable cells were counted using Trypan blue staining. *CD14⁺ versus CD14⁻: P < 0·05 at 24 hr, P < 0·01 at 48 hr, P < 0·0001 at 72 hr; †CD14⁺ versus CD14⁻/CD14CK: P < 0·01 at 72 hr, not significant at 24 and 48 hr; ‡CD14⁻ versus CD14⁻/CD14CK: P < 0·05 at 48 hr, P < 0·01 at 72 hr, not significant at 24 hr. n = 7, analysis of variance (anova). Bars show the mean value±1 SE. (e) Culture supernatants from CD14⁺ PBMC, CD14⁻ PBMC and CD14⁻ PBMC with addition of 20% CD14 cocktails were collected at 24, 48 and 72 hr, aliquoted and stored at −80°. IL-2 production was tested by enzyme-linked immunosorbent assay (ELISA). *CD14⁺ versus CD14⁻: P < 0·001 at 48 hr, P < 0·01 at 72 hr, not significant at 24 hr; †CD14⁺ versus CD14⁻/CD14CK: P < 0·05 at 48 and 72 hr, not significant at 24 hr. ‡CD14⁻ versus CD14⁻/CD14CK: P < 0·05 at 48 and 72 hr, not significant at 24 hr; n = 8, anova. Bars show the mean value±1 SE. (Panels (c), (d) and (e) are on the following page)

The CD14 cocktails were then tested for their ability to protect activated lymphocytes from undergoing AICD. First different doses of CD14 cocktails were added into CD14⁻ PBMC to test the protective effect. As shown in Fig. 3(b), CD14 cocktails successfully protected activated lymphocytes from undergoing AICD. However, no significant increase of the protective effect was observed with addition of an increasing amount (vol/vol) of CD14 cocktails. A 10% concentration of CD14 cocktails appeared to be sufficient to achieve this effect. To facilitate further comparison, a fixed amount (20%, vol/vol) was used throughout the experiments. Figure 3(c) shows that the addition of CD14 cocktails (CD14⁻/CD14CK) reversed the increased AICD rates in CD14⁻ PBMC at 24, 48 and 72 hr as compared with CD14⁺ PBMC (P < 0·01 at all time-points: CD14⁻ versus CD14⁻/CD14CK).

The effect of CD14 cocktails on lymphocyte growth was also evaluated. As shown in Fig. 3(d), addition of 20% CD14 cocktails at time 0 significantly increased the viable cell counts in the absence of CD14⁺ cells as compared with that in CD14⁻ PBMC (CD14⁺ versus CD14⁻: P < 0·05 at 24 hr, P < 0·01 at 48 hr, P < 0·0001 at 72 hr; CD14⁺ versus CD14⁻/CD14CK: P < 0·01 at 72 hr, it is not significant at 24 and 48 hr; CD14⁻ versus CD14⁻/CD14CK: P < 0·05 at 48 hr, P < 0·01 at 72 hr, it is not significant at 24 hr; n = 7, anova). These data suggested that the CD14 cocktails could promote the growth of activated lymphocytes. Furthermore, addition of CD14 cocktails until 24 and 48 hr had delayed effects on AICD and lymphocyte counts (data not shown).

To test if the CD14 cocktails could reverse the effects of CD14⁺ cell depletion on IL-2 secretion, the supernatants from PHA-p-activated lymphocytes were collected and tested for their IL-2 concentration using ELISA. As shown in Fig. 3(e), addition of CD14 cocktails (CD14⁻/CD14CK) significantly reversed the increased IL-2 secretion in the absence of CD14⁺ cells at 48 and 72 hr as compared with PHA-p-activated PBMC in the presence of CD14⁺ cells (CD14⁺ versus CD14⁻: P < 0·001 at 48 hr, P <0·01 at 72 hr, it was not significant at 24 hr; CD14⁺ versus CD14⁻/CD14CK: P < 0·05 at 48 and 72 hr, it was not significant at 24 hr; CD14⁻ versus CD14⁻/CD14CK: P < 0·05 at 48 and 72 hr, it was not significant at 24 hr; n = 8, anova).

A role of IL-6 in the protective effects of the CD14 cocktails

To characterize the active factors in the CD14 cocktails, we tested for common cytokines secreted by monocytes. IL-2 secretion was also tested for to exclude possible contamination of cytokines from activated lymphocytes. As shown in Fig. 3(a), the CD14 cocktails contained IL-1β, TNF-α, TGF-β and IL-6, but not IL-2, IL-12 and IL-15. The data suggested no contamination of cytokines from lymphocytes. Before analysing the possible role of these cytokines in the protective effects of the CD14 cocktails, 0·2 µm polyvinylidene fluoride (PVDF) membrane filters were used to filter the CD14 cocktails to exclude the possibility that molecules on cell membrane fragments were involved in the protective effects of the CD14 cocktails. Filtration did not affect the protective effect of the CD14 cocktails (data not shown). The CD14 cocktails were then filtered each time before use. TGF-β, as another candidate, was also tested. Owing to the finding that the media (10% FBS) used for culturing both CD14⁺ and CD14⁻ PBMC contained >1·0 ng/ml TGF-β (data not shown), it is highly unlikely that TGF-β is an active factor in the CD14 cocktails. The CD14 cocktails did not contain IL-12 and IL-15, as shown in Fig. 3(a).

In order to study other individual cytokines, depletion and blocking assays were designed as described in the Materials and methods. IL-1β and TNF-α were first tested as possible active factors in the CD14 cocktails. As shown in Fig. 3(a), the average IL-1β concentration was 293±33 pg/ml. The average TNF-α concentration was 1204±509 pg/ml. To test if either of these two cytokines contributed to the protective effect of the CD14 cocktails, they were depleted using R & D Systems' Quantikine ELISA kits (IL-1β and TNF-α). After three rounds of depletion, IL-1β and TNF-α levels were decreased by ≈ 85%. At this point, the residual IL-1 in the CD14 cocktail was 48±5 pg/ml and the residual TNF-α in the CD14 cocktail was ≈ 181±76 pg/ml. As 20% CD14 cocktails were used for the protection assays, the CD14 cocktails used for the protection assays still contained 10±1 pg/ml of IL-1β and 36±15 pg/ml of TNF-α. The residual IL-1β and TNF-α in the IL-1β- and TNF-α-depleted CD14 cocktails were further suppressed by adding the corresponding blocking mAbs. The corresponding blocking antibodies were added and the CD14 cocktails were further incubated at 37° for a further 1 hr (according to the protocol of R & D Systems). The data suggested that both TNF-α and IL-β were not the active factors in the CD14 cocktails (data not shown).

The same procedures as the analysis of IL-1β and TNF-α were then used to evaluate the possible role of IL-6 in the protective effects of the CD14 cocktails. As shown in Fig. 4(a), depletion by panning followed by blocking the residual IL-6 with anti-IL-6 mAb significantly decreased the protective effect of the CD14 cocktails, especially at the 72-hr time point during the 3-day culture period (CD14⁻ versus CD14⁻/CD14CK/IL-6D: P < 0·01 at 72 hr, it is not significant at 24 and 48 hr; n = 5, anova). To further prove the accuracy of this phenomenon, human recombinant IL-6 was used to determine its role in protecting activated lymphocytes from AICD. CD14⁺ and CD14⁻ PBMC were prepared as described above and then human recombinant IL-6 was added into CD14⁻ PBMC at concentrations of 0·5, 1, 2 and 4 ng/ml. As shown in Fig. 4(b), the addition of human recombinant IL-6 reversed the effects of CD14⁺ cell depletion on AICD, especially at 72 hr during the 3-day culture period (CD14⁻/IL-6/0·5 ng, CD14⁻/IL-6/1 ng, CD14⁻/IL-6/2 ng and CD14⁻/IL-6/4 ng) (CD14⁺ versus CD14⁻: P < 0·05 at 24, 48 and 72 hr; CD14⁺ versus CD14⁻/IL-6/0·5 ng, CD14⁻/IL-6/1 ng, CD14⁻/IL-6/2 ng and CD14⁻/IL-6/4 ng: P < 0·05 at 24 and 48 hr, it is not significant at 72 hr; CD14⁻ versus CD14⁻/IL-6/1 ng, CD14⁻/IL-6/2 ng and CD14⁻/IL-6/4 ng: P < 0·05 at 72 hr, it is not significant at 24 and 48 hr; n = 5, anova). The effect of rIL-6 on lymphocyte growth was also evaluated. As shown in Fig. 4(c), the lymphocyte counts were also increased at 24, 48 and 72 hr with the addition of recombinant IL-6 at concentrations of 0·5, 1, 2 and 4 ng/ml in the absence of CD14⁺ cells as compared with CD14⁻ PBMC alone (CD14⁻) (CD14⁺ versus CD14⁻: P < 0·05 at 24, 48 and 72 hr; CD14⁺ versus CD14⁻/IL-6 at 72 hr; P < 0·05 at all IL-6 concentrations except 4 ng/ml. CD14⁻ versus CD14⁻/IL-6 at 72 hr, P < 0·01 at an IL-6 concentration of 4 ng/ml, not significant at lower IL-6 concentrations; n = 4, anova).

Figure 4.

A role of interleukin-6 (IL-6) in acting as the survival signal in the CD14 cocktails. (a) IL-6 in the CD14 cocktails was depleted by panning, as described in the Materials and methods. The original CD14 cocktails and the CD14 cocktails depleted of IL-6 by panning followed by blocking with anti-IL-6 monoclonal antibody (mAb) were used for the protection assay. *CD14⁻ versus CD14⁻/CD14CK/IL-6D: P < 0·01 at 72 hr, not significant at 24 and 48 hr. n = 5, analysis of variance (anova). Bars show the mean value±1 SE. (b) CD14⁺ and CD14⁻ peripheral blood mononuclear cells (PBMC) were prepared as described in the Materials and methods. To CD14⁻ PBMC was added human recombinant IL-6 at concentrations of 0·5, 1, 2 and 4 ng/ml. CD14⁺ PBMC, CD14⁻ PBMC and CD14⁻ PBMC with different concentrations of IL-6 were stimulated with phytohaemagglutinin-p (PHA-p) at 37° in 5% CO₂. At 24, 48 and 72 hr, the activation-induced cell death (AICD) was analysed by flow cytometry using propidium iodide (PI)/annexin staining. *CD14⁺ versus CD14⁻: P < 0·05 at 24, 48 and 72 hr; †CD14⁺ versus CD14⁻/IL-6/0·5 ng, CD14⁻/IL-6/1 ng, CD14⁻/IL-6/2 ng and CD14⁻/IL-6/4 ng: P < 0·05 at 24 and 48 hr, not significant at 72 hr; ‡CD14⁻ versus CD14⁻/IL-6/1 ng, CD14⁻/IL-6/2 ng and CD14⁻/IL-6/4 ng: P < 0·05 at 72 hr, not significant at 24 and 48 hr; n = 5, anova. Bars show the mean value±1 SE. (c) To CD14⁻ PBMC were added human recombinant IL-6 at concentrations of 0·5, 1, 2 and 4 ng/ml. CD14⁺ PBMC, CD14⁻ PBMC and CD14⁻ PBMC with different concentrations of IL-6 were stimulated with PHA-p at 37° in 5% CO₂. At 24, 48 and 72 hr, the viable lymphocytes were counted in each well using the viable excluding dye Trypan blue. *CD14⁺ versus CD14⁻: P < 0·05 at 24, 48 and 72 hr; †CD14⁺ versus CD14⁻/IL-6 at 72 hr, P < 0·05 at all IL-6 concentrations except 4 ng/ml. ‡CD14⁻ versus CD14⁻/IL-6 at 72 hr, P < 0·01 at an IL-6 concentration of 4 ng/ml, it is not significant at other, lower, IL-6 concentrations. n = 4, anova. Bars show the mean value±1 SE. (d) IL-6 was added into CD14⁻ PBMC at concentrations of 0·5, 1, 2 and 4 ng/ml. CD14⁺ PBMC, CD14⁻ PBMC and CD14⁻ PBMC with IL-6 at different concentrations were cultured at 37° in 5% CO₂. At 24, 48 and 72 hr time-points, the supernatants were collected and tested for IL-2 concentrations by enzyme-linked immunosorbent assay (ELISA). The data represents the mean of three independent experiments. *CD14⁺ versus CD14⁻: P < 0·0001 at 48 and 72 hr; it is not significant at 24 hr; †CD14⁺ versus CD14⁻/IL-6: P < 0·01 at 48 and 72 hr time-points at all IL-6 concentrations; it is not significant at 24 hr. ‡CD14⁻ versus CD14⁻/IL-6: P < 0·05 at 72 hr at IL-6 concentrations of 2 and 4 ng/ml, it is not significant at other time-points and other, lower, IL-6 concentrations. n = 3, anova.

To determine whether the human recombinant IL-6 was working in a similar or the same manner as the CD14 cocktails, the IL-2 secretion in the presence or absence of IL-6 was tested using ELISA. As shown in Fig. 4(d), supernatants in the cultures contained a higher level of IL-2 in the absence of CD14⁺ cells as compared with that in the presence of CD14⁺ cells. Addition of human IL-6 at concentrations of 0·5, 1, 2 and 4 ng/ml partially reversed the effects of CD14⁺ cell depletion on IL-2 secretion (CD14⁻/IL-6/0·5 ng, CD14⁻/IL-6/1 ng, CD14⁻/IL-6/2 ng and CD14⁻/IL-6/4 ng) (CD14⁺ versus CD14⁻: P < 0·0001 at 48 and 72 hr; it is not significant at 24 hr; CD14⁺ versus CD14⁻/IL-6: P < 0·01 at 48 and 72 hr at all IL-6 concentrations; it is not significant at 24 hr. CD14⁻ versus CD14⁻/IL-6: P < 0·05 at 72 hr, at IL-6 concentrations of 2 and 4 ng/ml; it is not significant at other time-points and lower IL-6 concentrations; n = 3, anova).

Discussion

The immune response usually occurs within an environment tightly controlled by APC and lymphocytes. T-cell activation requires two signals.^36,37 However, the mechanisms for immune memory and immune tolerance are still not known. The survival of activated lymphocytes and/or other regulatory T cells is required for maintaining a normal immune response and immune memory. However, in genetically predisposed individuals, the survival and expansion of autoreactive lymphocytes and/or other regulatory T cells could result in immune disorders.^7–9,25 The death of activated lymphocytes is required for maintaining homeostasis of the immune system and removing harmful lymphocytes (autoreactive lymphocytes and other regulatory lymphocytes).^1–6,25 However, extensive death of lymphocytes could also result in immunodeficiency (e.g. as a result of human immunodeficiency virus and other infectious diseases).^41–45 We have postulated that T lymphocytes are activated upon receiving two signals (antigen and a costimulatory signal). At the same time, T cells receive a soluble survival signal from APCs. Only when activated T lymphocytes receive the survival signal, do the activated lymphocytes have sufficient time to migrate to remote sites, performing an immune function and generating immune memory.

To investigate this hypothesis, PHA-p was used to activate lymphocytes in the presence or absence of CD14⁺ cells. PHA-p binds CD2 (sheep red blood cell receptor) and activates T lymphocytes via the CD2 molecule.⁴⁶ The mechanisms of T lymphocyte activation by PHA-p have been previously investigated. Rosenstreich et al. reported absolute macrophage dependency of T-lymphocyte activation by mitogens.⁴⁷ Two different mechanisms were proposed, according to their findings. The first mechanism involves the binding of PHA to the macrophages followed by ‘presentation’ of the mitogen to the T lymphocytes in a manner that induces cell activation. The second mechanism is the elaboration of soluble factors by macrophages. The presence of these factors was demonstrated by using a double-chambered, Marbrook-type tissue culture vessel.⁴⁷ Further report supported that purified T11TS [a sheep form of lymphocyte function-associated antigen-3 (LFA-3)] could replace macrophages to help T-lymphocyte activation by PHA-p, demonstrating that provision of LFA-3 is a sufficient accessory cell function in the activation of human T cells by PHA-p.⁴⁸ Even though the CD2 molecule has its own intracellular signal transduction chain, T3-mutant E6-1 cells did not produce IL-2 in response to PHA-p, suggesting that the two pathways of T-lymphocyte activation (CD3, CD2) might be closely linked.⁴⁹

The data strongly suggested that CD14⁺ cells were important in preventing PHA-p-induced AICD, and that the AICD was a process of apoptosis, as determined by PI/annexin V staining,^50,51 DNA ladder⁵² and electron microscopy (EM) analysis.⁵³ Lymphocyte activation was delayed in the absence, as compared with that in the presence, of CD14⁺ cells, according to CD69 and CD25 expression. Lymphocyte growth was also decreased in the absence of CD14⁺ cells, according to cell counts in the same cultures. Analysis of T-lymphocyte subsets (CD3, CD4, CD8, CD3/4, CD3/8) also suggested that T lymphocytes were activated and underwent proliferation in both CD14⁺ and CD14⁻ PBMC. In addition, we found that supernatants from activated lymphocytes in the absence of CD14⁺ cells contained higher levels of IL-2 as compared with that in the presence of CD14⁺ cells. The increased IL-2 may come from the release of intracellular IL-2 as a result of the increased AICD or from increased transcription or translation of IL-2 in each cell.

To examine whether this protective effect was caused by the direct contact of the CD14⁺ cells or a resulting soluble factor, Dynabeads were used to activate CD14⁺ cells and induce secretion of monocytic cytokines (CD14 cocktails). The data suggested that Dynabeads activated CD14⁺ cells and induced secretion of IL-1β, TNF-α, IL-6 and TGF-β, which are common monocytic cytokines.^54,55 No IL-2 was found, suggesting that Dynabeads specifically activated CD14⁺ cells and induced secretion of monocytic cytokines. The Dynabeads were removed after incubation by using the Magnet Particle Concentrator (MPC). Therefore, the CD14 cocktails used here were not contaminated with any chemicals. The CD14 cocktails were then tested for their protective effect on AICD. These CD14 cocktails successfully prevented activated lymphocytes from undergoing AICD. This data reduced the possibility that CD14⁺ cells were only clearing apoptotic bodies and suggested a possible role of the soluble factor(s) produced by the activated CD14⁺ cells. The CD14 cocktails also reversed the elevated IL-2 production and decreased lymphocyte growth in the absence of CD14⁺ cells. In addition, the CD14 cocktails from serum-free medium had the same effects as that from RPMI-1640 culture medium, suggesting that bovine serum and 2-mercaptoethanol did not contribute to the observed effects. In addition, delayed addition of CD14 cocktails until 24 and 48 hr had delayed effects on AICD and lymphocyte growth. The data also suggested that the CD14 cocktails could rescue activated lymphocytes from undergoing AICD. However, data from other laboratories have shown no effect on apoptosis of soluble factors from activated monocytes.^28,29 This might be, in part, a result of the different systems used. The previously used system was a classic two-chamber system. It was found, in the presence of activated T lymphocytes and monocytes in the upper chamber, that T lymphocytes in the lower chamber were not rescued from anti-CD-induced AICD. Pretreatment of activated T lymphocytes with monocytes for 24 hr, or in the presence of monocytes, rescued T lymphocytes from AICD upon anti-CD3 activation. However, the experiments did not give evidence showing that the monocytes in the presence of activated T cells in the upper chamber were actually secreting soluble factors and/or that the soluble factors from the upper chamber were present in the lower chamber in this system.

The active working factors were also investigated. First, to exclude the possible role of membrane fragments (CD40/CD40L, CD80/CD86 and CD95L) in the CD14 cocktails,⁵⁶ CD14 cocktails were filtered through 0·2 µm PVDF membranes. The results suggested that membrane fragments could not account for the protective effect in the CD14 cocktails. By using a panning method to deplete individual cytokines, we were able to deplete more than 85% of the specific cytokines. After depletion, the residual cytokines in the CD14 cocktails were further blocked by corresponding mAbs. The data suggested that neither IL-1β nor TNF-α was the active protective factor in the CD14 cocktails. TGF-β was not the active factor that altered the AICD rates in the CD14⁺ and CD14⁻ PBMC, because TGF-β was already present at >1·0 ng/ml in the media used for culturing both CD14⁺ and CD14⁻ PBMC. The above data is consistent with a previous report showing that the presence of a significant amount of TGF-β in bovine serum accounted for the presence of TGF-β in general culture media if bovine serum was used.⁵⁷ Importantly, it was further found that depletion of IL-6 by panning followed by blocking residual IL-6 with anti-IL-6 mAb significantly decreased the protective effect of the CD14 cocktails, especially at 72 hr, suggesting that IL-6 may partially account for the protective effect of the CD14 cocktails. By using human recombinant IL-6, the protective effect of human recombinant IL-6 on AICD was shown, especially at the 72 hr time-point. Further analysis showed that human recombinant IL-6 also partially reversed the effect of CD14⁺ cell depletion on lymphocyte growth and IL-2 secretion, suggesting that the effect of human recombinant IL-6 was similar to that of the CD14 cocktails. However, the activity of the CD14 cocktails was not fully restored by recombinant IL-6. This data implied that some other, as yet unidentified, additive factors exist in the CD14 cocktails.

IL-6 belongs to a family including IL-11, LIF, OSM, CNTF and CT-1, in which all the cytokines use a common receptor subunit (gp130) for signal transduction.^58,59 Early in 1992, it was found that large proportion of T cells isolated from patients with acute infectious mononucleosis died rapidly, compared with only a few T cells from normal individuals. Interestingly, addition of rIL-6 was able to rescue those apoptosis-sensitive T lymphocytes from undergoing apoptosis.⁶⁰ After that, several groups reported that IL-6 could also inhibit tumour suppressor p53 protein-induced apoptosis in M1 myeloid leukaemic cells^61,62 T-cell receptor (TCR)-mediated apoptosis of neonatal T cells,⁶³ TCR/CD3-mediated apoptosis of T-cell hybridoma¹⁹ and spontaneous apoptosis of explanted lymph node/splenic resting CD4⁺ T cells.⁶⁴ In the latter observation, IL-6 appeared to induce or maintain Bcl-2 expression in explanted CD4⁺ resting T cells.⁶⁴ Our data suggested that anti-CD14-activated monocytes could secrete IL-6, which acted as one of the active survival factors in the CD14 cocktails. The importance of this finding resides in its integration of three signals into one entity – APC. The three signals are antigens, costimulatory molecules and soluble survival factors. Ongoing investigations are studying the role of the proposed soluble survival signal in maintaining pathogenic T lymphocytes in rheumatoid arthritis, in the survival of other T lymphocytes and in the generation of immune memory by using anti-CD3 activation and/or other antigen-specific T-cell lines.

Based on these findings, we propose a soluble survival signal hypothesis for lymphocyte activation and proliferation. In normal situations, this survival signal allows activated lymphocytes to stay alive, maintaining an effective immune response and generating immune memory. However, a disorder of this survival signal could result in either autoimmune diseases or immunodeficiency in genetically predisposed individuals. The above data integrate lymphocyte activation, AICD, peripheral self-tolerance, immune memory and immune regulation. It also provides an important contribution to our current knowledge about the basic mechanisms of the immune response.

Abbreviations

CD14⁺ PBMC (or CD14⁺)

PBMC before CD14⁺ cell depletion

CD14⁻ PBMC (or CD14⁻)

PBMC after CD14⁺ cell depletion

CD14⁺ cells

Dynabead-bound cells

CD14 cocktails

supernatants from cultures of Dynabead-bound cells (CD14⁺ cells)

CD14⁻/CD14CK

CD14⁻ PMBC with addition of CD14 cocktails

CD14⁻/CD14CK/IL-6D

CD14⁻ PBMC with addition of CD14 cocktails that were depleted of IL-6 and followed by blocking with anti-IL-6 mAb

CD14⁻/IL-6/0·5 ng, CD14⁻/IL-6/1 ng, CD14⁻/IL-6/2 ng, CD14⁻/IL-6/4 ng, CD14⁻

PBMC with addition of recombinant human IL-6 at concentrations of 0·5, 1, 2 and 4 ng/ml, respectively

EAE

experimental allergic encephalomyelitis