Levels of expression of complement regulatory proteins CD46, CD55 and CD59 on resting and activated human peripheral blood leucocytes

Abstract

The cell surface complement regulatory (CReg) proteins CD46, CD55 and CD59 are widely expressed on human lymphoid and non-lymphoid cells. This study aimed to compare systematically levels of CReg expression by different leucocyte subsets and to determine whether levels were increased following activation in vitro. Levels of each CReg protein were similar on freshly isolated monocytes and all major lymphocyte subsets, except that CD4⁺ cells expressed significantly less CD46 than CD8⁺ cells (P < 0·05) while the reverse was observed for CD55 (P < 0·02). CD56⁺ cells, predominantly natural killer cells, expressed significantly lower levels of CD59 than T cells (P < 0·02). CD45RO⁺ cells had higher levels of surface CD46 and CD59, but lower levels of CD55, than CD45RO^– cells (P < 0·02); CD25⁺ cells also expressed significantly less CD55 than CD25⁻ cells (P < 0·002). Neutrophils expressed higher levels of CD59, but lower levels of CD55, than monocytes. Following activation with phytohaemagglutinin, CD46 was up-regulated on all leucocyte subsets with the exception of CD56⁺ cells. Both CD55 and CD59 were also markedly up-regulated on monocytes, and CD55 expression was greater on CD8⁺ than CD4⁺ cells following activation (P < 0·02). Lipopolysaccharide treatment did not significantly alter B-cell expression of CReg proteins whereas CD55 and CD59, but not CD46, were significantly up-regulated on monocytes (P < 0·02). These observations that CReg proteins are up-regulated on certain activated leucocyte subsets indicate that levels would be increased following immune responses in vivo. This could enhance both protection against local complement activation at inflammatory sites and also the immunoregulatory properties of these leucocytes.

Introduction

Complement regulatory (CReg) proteins are widely expressed across many cell lineages and their best characterised function is to protect cells from autologous complement-mediated lysis. There are three major human cell surface CReg proteins: CD46 (membrane cofactor protein) facilitates C3b and C4b inactivation, CD55 (decay accelerating factor) disrupts C3 and C5 convertases, and CD59 prevents membrane attack complex assembly.¹

In addition to their function in complement regulation, each of these CReg proteins has been reported to have other immune functional activities. CD46 can act as a costimulatory molecule in human CD3-dependent T-cell activation and its engagement induces morphological changes and actin reorientation.² In a transgenic mouse model, human CD46 was able to regulate T-cell dependent inflammatory responses.³ Parallel triggering of T cells with anti-CD3 and anti-CD46, or the streptococcal M protein ligand for CD46, can induce interleukin-10 (IL-10)-producing T cells with regulatory capacity.^4,5 In human monocytes, triggering of CD46 induces IL-12 p40 and nitric oxide production as well as recruitment of src homology 2 domain-containing phosphatase-1 (SHP-1).⁶ It has been shown also that CD46 is lost rapidly on apoptotic and necrotic cells, facilitating their rapid complement-mediated removal.⁷

Experiments in mice have demonstrated that the glycosylphosphatidylinositol (GPI)-linked CD55 protein acts as a negative modulator of T-cell responses.^8,9 CD55 can act as an intercellular adhesion ligand, interacting with CD97 on leucocytes¹⁰ and CD55 expressed on mucosal epithelial cells can also regulate neutrophil movement across epithelial monolayers.¹¹ CD55 has been particularly studied on neutrophils, where it is dynamic and can be endocytosed and shed.¹² Variations in CD55 levels have been described in neutrophils following apoptosis and activation.^13,14 Engagement of keratinocyte CD59 can lead to signal transduction and nuclear factor-κB (NFκB)-mediated triggering of cytokine production.^15,16 CD59 on natural killer (NK) cells plays a role as a costimulatory molecule in mediating cytotoxic function¹⁷ and is involved in adhesion and subsequent activation of human T cells.¹⁸ The GPI-linked form of CD59 increases target cell susceptibility to NK cell-mediated lysis.¹⁹ When expressed on neutrophils, CD59 can interact with extracellular calreticulin and mediate signal transduction following calreticulin cross-linking.²⁰ However, when expressed on T cells, CD59 cross-linking can induce cell death.²¹

The aims of the present study were to compare systematically the levels of CReg protein expression on resting peripheral blood leucocyte subpopulations and to determine whether levels of cell surface expression change following cell activation in vitro.

Materials and methods

Cells

Peripheral blood mononuclear cells (PBMC) were obtained from a panel of healthy unrelated subjects (12 males, 14 females; age range 22–52 years) with informed consent. Ethical permission for the study was obtained from the Liverpool Research Ethics Committee. Blood was anticoagulated with heparin and mononuclear cells purified by density gradient centrifugation on Ficoll-Paque™ Plus (Amersham Biosciences, Amersham, UK). After washing in phosphate-buffered isotonic saline (PBS), cells were resuspended at 10⁶/ml in culture medium (CM) comprising RPMI-1640 with 10% heat-inactivated fetal calf serum (FCS), 2 mm l-glutamine, 50 U/ml penicillin and 50 µg/ml streptomycin. The human cell lines Jurkat (T-cell leukaemia) and U937 (promonocyte) were obtained from European Collection of Animal Cell Cultures (Porton Down, UK) and used as positive controls for CReg protein expression.

Cell activation

In some experiments, PBMC were incubated for 2 days at 37° in an atmosphere containing 5% CO₂ with CM alone, with 1 µg/ml phytohaemagglutinin (PHA; Sigma, Poole, UK) or with 100 ng/ml lipopolysaccharide (LPS; Sigma-Aldrich). After culture for 4 hr, 18 hr or 2 days, flasks were agitated gently to remove loosely adherent cells before analysis.

Antibodies

The following conjugated test or isotype-control monoclonal antibodies were used against CReg proteins: anti-CD46–fluoroscein isothiocyanate (FITC), anti-CD55–FITC, anti-CD59–FITC (BD Pharmingen, San Diego, CA); against leucocyte subsets: anti-CD3–phycoerythrin (PE, T cells; Dako, High Wycombe, UK), anti-CD4–PE, anti-CD8–PE, anti-CD19–PE (B cells; Sigma-Aldrich), anti-CD14–PE (monocytes), anti-CD56–PE (NK cells and some T cells; Caltag (Invitrogen), Paisley, UK), anti-CD25–PE Cy5 (IL-2Rα, regulatory and activated T cells) and anti-CD45RO–PE Cy5 (recently activated/memory T cells; BD Pharmingen); isotype-controls: IgG1-PE (Immunotech, Fullerton, CA) and IgG2a-PE (Caltag).

Flow cytometry

Freshly isolated cells, or cells that had been cultured for 2 days with or without PHA or lipopolysaccharide (LPS), were washed in PBS, resuspended and 50 µl aliquots (5 × 10⁵ cells) placed into 5 ml plastic tubes. Monoclonal antibodies (2 µl) were added and the cells incubated for 30 min at 4°. Cells were then washed twice in cold PBS and resuspended in 0·5 ml PBS for analysis. Cells were incubated with either isotype-matched control antibodies or a combination of two antibodies, an FITC-labelled antibody against one of the CReg proteins and a PE-labelled antibody against a human leucocyte subset. Control Jurkat and U937 cells were labelled in the same way as PBMC.

In some experiments, heparinized whole blood was labelled directly without exposing leucocytes to Ficoll-Paque™. Aliquots of 50 µl blood were labelled with 5 µl of each antibody conjugate for 10 min at room temperature, followed by erythrocyte lysis, cell stabilization and cell fixation using an automated system with associated reagents (TQ-Prep Workstation; Beckman Coulter, Fullerton, CA). Mononuclear cells from the same blood sample were prepared in parallel, labelled with the same antibody concentrations and analysed sequentially.

Cells were analysed using an EPICS XL MCL flow cytometer (Coulter) equipped with proprietary software. Lymphocytes, monocytes or neutrophils, as appropriate, were gated on the basis of forward and side scatter. Cells labelled with isotype-controls were used to set the fluorescence thresholds for positivity such that <1% of cells labelled with isotype-control antibodies were positive. Levels of expression of CReg proteins on leucocyte subsets were assessed based on the basis of dual staining with FITC–anti-CReg protein and PE–anti-leucocyte subset antibodies. Mean fluorescence intensities (MFI) were recorded for CReg protein labelling of cells of each leucocyte subpopulation cultured with and without stimulation. Relative changes in expression following stimulation were expressed as percentage MFI compared to unstimulated cells cultured in CM alone.

Statistical analysis

Student's t-test (2-tailed) was used to compare differences in levels of CReg protein expression between leucocyte subsets and between activated and nonactivated cells; P-values <0·05 were regarded as statistically significant.

Results

Relative levels of CReg protein expression on leucocyte subsets

Relative levels of expression, assessed by mean fluorescence intensities, were measured on all major mononuclear cell subsets in freshly isolated PBMC from eight normal subjects. Both control human cell lines expressed all three CReg proteins, with U937 cells expressing particularly high levels of CD46, while Jurkat cells strongly expressed CD59 (Fig. 1). All PBMC subsets expressed the three CReg proteins, although mean levels were significantly different in some instances. CD4⁺ cells expressed significantly lower levels of CD46 than CD8⁺ cells (Fig. 1a;P < 0·05), while the opposite was found for CD55 (Fig. 1b;P < 0·02). CD56⁺ cells, predominantly NK cells, expressed markedly lower levels of CD59 than T cells (Fig. 1c;P < 0·02).

Figure 1.

Expression of (a) CD46 (b) CD55 and (c) CD59 by control human cell lines (Jurkat and U937) and leucocyte subsets as indicated (PMN = neutrophils). Results are expressed as means of MFI⁺-standard deviation (SD). (a) *CD4⁺ cells versus CD8⁺ cells (P < 0·05); (b) *CD8⁺ cells versus CD4⁺ cells (P < 0·02), ††PMN versus monocytes (P < 0·001); (c) *CD56⁺ cells versus CD3⁺ cells (P < 0·02), †PMN versus monocytes (P < 0·03). No other differences between PBMC subsets were statistically significant.

CReg protein expression was compared also on freshly isolated lymphocytes stained for expression of the IL-2 receptor α-chain CD25 (activated/regulatory T cells) and the recently activated/memory marker CD45RO. Levels of CD55 were significantly lower on both CD25⁺ and CD45RO⁺ cells than the respective negative populations (Table 1). Also, both CD46 and CD59 were expressed at higher levels on CD45RO⁺ than CD45RO^– cells (Table 1), but no other differences were statistically significant.

Table 1.

CR protein expression by freshly isolated lymphocyte subsets (CD25^± and CD45RO^±)

CR protein	Cell subpopulation	MFI (arbitrary units) ± SD	P-value
CD46	CD25+	46·9 ± 7·5	0·5
	CD25−	48·7 ± 4·4
CD55	CD25+	24·6 ± 2·8	<0·002*
	CD25−	40·4 ± 4·5
CD59	CD25+	49·6 ± 3·7	0·08
	CD25−	44·8 ± 7·4
CD46	CD45RO+	32·9 ± 9·6	0·006*
	CD45RO−	24·9 ± 6·5
CD55	CD45RO+	17·0 ± 4·4	0·02*
	CD45RO−	22·2 ± 6·9
CD59	CD45RO+	33·3 ± 11·2	0·01*
	CD45RO−	26·2 ± 8·6

^*Differences between positive and negative subset statistically significant.

Effect of Ficoll separation on CReg protein expression

Leucocytes prepared using standard Ficoll density gradient separation were compared with those prepared from whole blood by erythrocyte lysis (n = 5). The latter technique permitted analysis of CReg protein expression by polymorphonuclear leucocytes (PMN; neutrophils), which were found to express significantly greater levels of CD59 than monocytes prepared in the same way (P < 0·03) but significantly lower levels of CD55 than monocytes (P < 0·001) (Fig. 2).

Figure 2.

Effect of preparing leucocytes using Ficoll-Paque™ on CReg protein expression. Monocytes (solid bars) and lymphocytes (hatched bars) from the same donor were analysed simultaneously following purification by erythrocyte lysis (WB) or using Ficoll-Paque™. Results are expressed as a percentage of CReg protein expression. *P < 0·02; **P < 0·01; ***P < 0·005.

When the same cell populations prepared using either Ficoll separation or erythrocyte lysis were compared, monocytes uniformly expressed higher levels of all three CReg proteins following Ficoll separation, but this was only statistically significant for CD46 (P < 0·02) and CD55 (P < 0·005) (Fig. 2). However, the converse was found for lymphocytes, with cells prepared directly from whole blood having significantly increased levels of all three CReg proteins (P < 0·01; Fig. 2).

Changes in CReg protein expression following PHA activation

Several significant increases in CReg protein expression were noted following PHA activation (n = 8), but the changes were not uniform on all cell subsets. As shown in Fig. 3(a), CD46 was significantly up-regulated on both CD4⁺ and CD8⁺ cells as well as on monocytes (P < 0·02), while no change in expression was noted for CD56⁺ cells. However, CD55 was significantly up-regulated only on CD8⁺ cells and monocytes (P < 0·02), while expression on CD4⁺ cells was unaffected (Fig. 3b). Monocytes also showed strong up-regulation of CD59 (P < 0·02), while lymphocyte subsets showed no change in expression (Fig. 3c). Cells cultured in CM alone for 48 hr showed similar levels of CReg protein expression to freshly isolated cells (data not shown). In further experiments (n =5), a time course study was performed using different times of incubation with PHA. In each case, significant changes in CReg protein expression were observed at 4 and 18 hr, but were less marked than those after 2 days (data not shown).

Figure 3.

Changes in MFI of (a) CD46 (b) CD55 and (c) CD59 expressed on T-cell subsets, NK cells and monocytes following activation by PHA. Filled bars represent cells cultured for 2 days in CM alone; hatched bars represent cells cultured with PHA for 2 days. Results are expressed as percentage change relative to unstimulated cells ± SD. *P < 0·02.

Changes in CReg protein expression following LPS activation

Following activation with LPS, no significant increases in levels of CD46 were observed on either on B cells or monocytes (Fig. 4). However, levels of CD55 (P < 0·01) and CD59 (P < 0·002) were significantly increased on monocytes but not B cells (Fig. 4).

Figure 4.

Changes in MFI of staining on B cells and monocytes for CD46 (filled bars), CD55 (open bars) and CD59 (hatched bars) following activation with LPS for 2 days. Increases in monocyte expression were statistically significant for CD55 (*P < 0·01) and CD59 (**P < 0·002).

Discussion

All freshly isolated human leucocyte subsets analysed in this study expressed each of the three CReg proteins, although some significant differences in levels of expression were noted for CD4⁺, CD8⁺ and CD56⁺ cells. This is in contrast to the findings of Terstappen et al. who identified CD55 only on a subset of both T cells and NK cells.²² The lower levels of CD55 (not statistically significant) and CD59 on CD56⁺ cells (predominantly NK cells) compared to T cells are in agreement with an earlier report of reduced expression in the predominant CD3^– CD56^dim peripheral NK cell subset.²³ In another report, NK cell levels of CD59 were lower, but not significantly so, than those on T cells.²⁴ While higher levels of CD59 were found on neutrophils than on monocytes, the opposite was observed for CD55. This, together with the strikingly different levels of expression by Jurkat and U937 cell lines, indicates that expression of CReg proteins is generally not coordinated.²⁵ The high levels of expression of CD46 on U937 cells, and of CD59 on Jurkat cells, may be an adaptation to long-term culture in vitro. In the latter case, this may explain the sensitivity of Jurkat cells to apoptosis following CD59 cross-linking.²¹

The differential changes in CReg protein expression following cell activation with PHA or LPS reinforce the concept that regulation of expression of the three CReg proteins is neither coordinated nor uniform in different leucocyte subsets.²⁵ Monocytes were the only cell type showing uniform up-regulation following treatment with PHA and, with the exception of CD46, LPS stimulation. This up-regulation by monocytes was additional to that resulting from the use of Ficoll to separate the cells, and hence is likely to be physiologically relevant. LPS acts via Toll-like receptor-4 (TLR4) to induce NFκB activation²⁶ and this may be the mechanism of up-regulation. Both CD55 transfected into Chinese hamster ovary cells²⁷ and CD59 on oral keratinocytes²⁸ can transduce signals following LPS stimulation. However, under the same conditions, B cells failed to up-regulate any of the CReg proteins, possibly because of a lack of TLR4 expression.

The precise mechanism of action of the lectin PHA in activating T cells is not certain but is likely to require the participation of accessory cells, such as dendritic cells or monocytes.²⁹ In the present experiments, PHA may be able to interact directly with monocytes and induce CReg protein up-regulation. This has been reported for up-regulation of CD55 expression on human endothelial cells stimulated with PHA and other lectins.³⁰ Alternatively, cytokines such as IL-2 or interferon-γ produced by T cells, following stimulation with PHA, might indirectly induce CReg protein up-regulation in monocytes. PHA induces CD25 expression in T cells³¹ and CD122 in monocytes,³² while LPS can induce CD25 on monocytes.³³ This would increase the sensitivity of T cells and monocytes to IL-2 produced following mitogenic stimulation.³⁴ Cytokines have been shown to up-regulate expression of CD55 and CD59, but not of CD46, on human endothelial cells^25,34 and fibroblasts.³⁵

The up-regulation of CReg proteins observed following non-specific activation of T cells and monocytes in vitro may also reflect changes taking place following antigen-specific activation in vivo. This would be expected to influence the immunoregulatory properties of T cells and monocytes as all three CReg proteins have been reported to have other immune functional activities when expressed on peripheral blood mononuclear cells.^{2–6,8–10,17–19,21} Future work could investigate changes in CReg protein expression in vitro or in vivo following antigen-specific immune responses. The involvement of cytokines in up-regulation of CReg protein expression might be studied on isolated cell populations to determine their relative importance as inducers of CReg protein expression.

The observations that CD46 and CD59, but not CD55, are expressed at higher levels on recently activated/memory lymphocytes (CD45RO⁺) than naïve lymphocytes (CD45RO^–) could indicate that the changes in CReg protein expression observed in vitro occur also following lymphocyte activation in vivo. However, the present analysis of the small subset of CD25⁺ cells would include both activated T cells and regulatory T cells. Future work could analyse CReg protein expression by better defined subsets of innate and activated regulatory T cells.

It is well known that CReg proteins expressed on many cell types can act as cellular receptors for a diverse range of microbial pathogens.³⁶ Any variations in CReg protein levels on different leucocyte subsets may lead to differential susceptibilities to infection as well as to complement-mediated damage. Hence, changes in CReg protein levels following leucocyte activation might alter susceptibility to both of these processes. In paroxysmal nocturnal haemoglobinuria, reduced levels of GPI-linked proteins, which include CD55 and CD59, result in enhanced susceptibility of erythrocytes to complement-mediated lysis.³⁷ Significantly reduced levels of CD59 are found also on all lymphocyte subsets, as well as on monocytes and neutrophils, in this disease.²³ Hence, alterations in CReg protein expression may have pathological consequences. This is illustrated also in studies in rheumatoid arthritis patients showing CD46 expression on synovial fluid neutrophils to be lower than on cells in matched peripheral blood, whereas levels of CD55 were higher.³⁸

In conclusion, our observations that CReg proteins are up-regulated on certain activated leucocyte subsets indicates that levels would be increased following immune responses in vivo, enhancing both protection against local complement activation at inflammatory sites and also the immunoregulatory properties of these leucocytes.

Abbreviations

CM

culture medium

Creg

complement regulatory

FCS

fetal calf serum

FITC

fluorescein isothiocyanate

GPI

glycosylphosphatidylinositol

IL

interleukin

LPS

lipopolysaccharide

MFI

mean fluorescence intensity

NK

natural killer

PBMC

peripheral blood mononuclear cells

PBS

phosphate-buffered isotonic saline

PE

phycoerythrin

PHA

phytohaemagglutinin

PMN

polymorphonuclear (leucocyte)

TLR

Toll-like receptor