Expression of caveolin by bovine lymphocytes and antigen-presenting cells

Abstract

Caveolin is a generic term for a family of proteins that include caveolin-1, -2 and -3. Although the distribution of these proteins varies between cells, caveolin-1 and -2 are commonly found coating membrane invaginations known as caveolae. Studies on human and murine cells suggest that caveolin/caveolae can be found in neutrophils, macrophages and mast cells, in which they are involved in the uptake of pathogens, but not in lymphocyte cell lines. Expression of caveolin-1, -2 and -3 in bovine immune cells was investigated using confocal microscopy and Western blotting. Staining for caveolin-1 was evident in all peripheral blood mononuclear cells (PBMC) and in CD4⁺, CD8⁺ and CD21⁺ lymphocytes, monocytes, macrophages and monocyte-derived dendritic cells (DC). In addition, the caveolin-1 antibody detected a protein with a molecular weight of approximately 22000 in all PBMC, macrophages and DC, as well as in bovine aortic endothelial (BAE)-1 cells and human endothelial cells by Western blotting. In macrophages and DC, caveolin co-localized with the endoplasmic reticulum–Golgi intermediate compartment (ERGIC) and to a lesser extent with Golgi, but not with endoplasmic reticulum. Staining was not seen on the plasma membrane in any bovine immune cells, suggesting the absence of caveolae, while in BAE-1 cells staining was predominantly on the cell membrane. Caveolin-2 could not be detected in any bovine cells by confocal microscopy or Western blotting, while caveolin-3 was detected in all bovine cells by Western blotting, but not by confocal microscopy. These data provide evidence for the presence of caveolin in bovine lymphocytes and antigen-presenting cells.

Introduction

Caveolae are small, flask-shaped invaginations found in cholesterol/glycosphingolipid-enriched microdomains of the plasma membrane and characterized by the presence of the protein caveolin. In many cells, caveolin is thought to be involved in fatty acid and cholesterol transport and cycles between the Golgi, endomplasmic reticulum (ER) and plasma membrane caveolae.^1,2 However, caveolae are not the major source of caveolin in all cells; in skeletal muscle cells and keratinocytes it is found in the cytosol, in exocrine and endocrine cells it is found in the secretory pathway, and in airway epithelial cells it accumulates in modified mitochondria.³ Three caveolin genes have so far been cloned; caveolin-1 (of which there are two isoforms, α and β),⁴ caveolin-2⁵ and caveolin-3.⁶ Although similar in structure and function, these caveolin isoforms differ in their cellular distribution and specific properties. Caveolin-1 and -2, which together form a stable hetero-oligomeric complex, are most abundant in endothelial cells, smooth muscle cells, skeletal myoblasts, fibroblasts and adipocytes,⁷ while the expression of caveolin-3 is thought to be largely muscle specific.^8,9

Caveolin has been detected in human T-cell leukaemia cell lines,¹⁰ murine macrophages and mast cells^11–13 and human and bovine dendritic cells.¹⁴ Conflicting evidence exists regarding the presence of caveolin in human neutrophils^15,16 and Fra et al.¹⁷ could not detect the protein or caveolae in murine and human T- and B-cell lines. Caveolae have been implicated in the internalization of FimH-expressing bacteria by mouse macrophages and mast cells^13,18 and of respiratory syncytial virus by bovine dendritic cells.¹⁴ Indeed, it has been suggested that a number of microbes or their exotoxins may exploit caveolae-dependent endocytosis as a means of ensuring their subsequent intracellular survival.¹⁹ In addition, activation of rat and mouse macrophages has been shown to induce caveolin expression.^11,12 Caveolin-1 mRNA expression is up-regulated by lipopolysaccharide in murine C3HeB/FeJ peritoneal macrophages,¹² while ultrastructural and functional studies have demonstrated that the number of caveolae on rat peritoneal macrophages is increased by elicitation with adjuvant in vivo.^20,21 In view of this growing body of evidence for the involvement of caveolae in the uptake of pathogens by antigen-presenting cells (APC), we have examined the distribution of caveolin in bovine peripheral blood mononuclear cells (PBMC) and show that it can be detected in both APC and lymphocytes.

Materials and methods

Animals

Blood was taken from healthy cattle (Bos taurus) from the Institute for Animal Health herd and the Faculty of Veterinary Medicine, University of Zurich, and placed in heparin (10 U/ml).

Antibodies

Rabbit polyclonal antibodies against caveolin-1 (N-20) and caveolin-2 (H-96) and goat polyclonal antibody against caveolin-3 (N-18) (Santa Cruz Biotechnology Inc., Santa Cruz, CA) were used for confocal analysis and Western blotting. A second rabbit polyclonal antibody against caveolin (C13630, BD Transduction Laboratories, Lexington, KY) was also tested. Mouse monoclonal antibodies (mAb) to the following leucocyte differentiation antigens were used; CD4 [mAb CC8, immunoglobulin G2a (IgG2a)], CD8 (mAb CC63, IgG2a), CD14 (mAb CCG33, IgG1) and CD21 (mAb CC21, IgG1), bovine IgM was detected with IL-A30 (mouse, IgG1).^14,22,23 Mouse mAb against golgin-97 (IgG1) was obtained from Molecular Probes Europe BV, Poortgebouw, the Netherlands, mouse mAb against protein disulphide isomerase (PDI, IgG1) from Bioquote Ltd, York, UK and G1/G93 (mouse IgG1) against human endoplasmic reticulum/Golgi intermediate compartment (ERGIC-53) was a generous gift from H-P. Hauri, Biozentrum, Basel, Switzerland. For controls, rabbit IgG was purified from normal rabbit serum and murine mAbs to chicken surface proteins (AV20 – Bu-1, IgG1; AV37 – chicken spleen cell subset, IgG2a), kindly provided by F. Davison, IAH, Compton, were used.

Preparation of cells

PBMC were separated by density gradient centrifugation on 1·083 g/ml Histopaque (Sigma, Poole, Dorset, UK) and cytospun (60×g, 10 min) onto glass slides. CD14⁺ cells were isolated from a Midimacs column after incubation with anti-human CD14-labelled super-paramagnetic beads (Miltenyi-Biotech, Bergisch Gladbach, Germany). CD4⁺, CD8⁺ and CD21⁺ cells were first labelled with specific mouse mAb, then incubated with anti-mouse IgG-labelled super-paramagnetic beads. The purity of the sorted cells, evaluated by flow cytometry, was > 98% in all experiments. Cell viability, evaluated by trypan blue exclusion, was > 95%. CD14⁺ cells were incubated on glass coverslips in 24-well plates at a concentration of 10⁶ cells/ml in RPMI-1640 culture medium with glutamax (Life Technologies Ltd, Paisley, Renfrewshire, UK) supplemented with 10% heat-inactivated fetal calf serum (FCS), 5 × 10⁻⁵ m 2-mercaptoethanol and 50 µg/ml gentamicin (tissue culture medium, TCM). Monocytes were harvested after 2 hr in culture, macrophages after 3 days. Alternatively, monocyte-derived dendritic cells (DC) were generated using a previously described method.²³ Briefly, CD14⁺ cells were grown on coverslips in TCM supplemented with 200 U/ml bovine recombinant interleukin-4 and 0·2 U/ml bovine recombinant granulocyte–macrophage colony-stimulating factor for 7 days. On day 3 of culture 0·5 ml of spent media was replaced with fresh TCM.

Bovine aortic endothelial (BAE-1) cells (obtained from the European collection of cell cultures, ECACC, 88031149) were cultured on glass coverslips in 24-well plates in Dulbecco's modified Eagle's minimal essential medium (DMEM) with 10% FCS and fetal calf kidney (FCK) cells cultured on coverslips in DMEM with 10% FCS, 5% lactalbumin hydrolysate, 1% non-essential amino acids, 1% HEPES, 5 ng/ml epidermal growth factor and 10000 U/ml penicillin/streptomycin.

Western blotting

BAE-1 cells and cells sorted from PBMC were treated with 0·1 ml of filtered (0·45 µm) lysis buffer consisting of 0·15 m NaCl, 0·05 m Tris–HCl (pH 7·2), 0·1% sodium dodecyl sulphate, 1% Triton-X-100, 1 mm phenylmethylsulphonyl fluoride and 1% sodium deoxycholate overnight on ice. The lysate was then centrifuged at 14000 g for 30 min. The supernatant was collected and kept at −20° for future use and the pellet was discarded. As a positive control, a human endothelial cell (HEC) lysate, derived from an aortic endothelium cell line (BD Transduction Laboratories) was used at a concentration of 0·1 mg/ml. The protein samples were heat-denatured in Laemmli buffer and cellular proteins (10 µg per lane) were separated on Protean II mini-gels (Bio-Rad, Hercules, CA) on a 16% sodium dodecyl sulphate–polyacrylamide gel electrophoresis, and electrically (100 V, 100 min) transferred to PVDF membranes. Thereafter, non-specific binding was blocked with TBST (50 mm Tris–HCl, pH 7·5, 0·15 m NaCl, 0·05% Tween-20) containing 5% fat-free milk powder overnight at 4°. Membranes were incubated with the first antibody (1:500) in the above blocking solution for 2 hr, washed in blocking solution, and incubated with the secondary antibody (diluted 1:5000 in blocking buffer) for 2 hr. After additional washing, bound antibodies were visualized using ECL reagents, followed by exposure to X-ray films.

Confocal microscopy

Cells were fixed in ice-cold methanol for 5 min, then treated with 0·5% bovine serum albumin in phosphate-buffered saline (PBS/BSA) to prevent non-specific binding and incubated with primary antibodies in PBS/BSA for 1 hr at room temperature. After washing three times in PBS, the cells were incubated with goat anti-rabbit or donkey anti-goat IgG Alexa-fluor 568 -conjugated antibody and goat anti-mouse IgG Alexa-Fluor 488 (1:200 in PBS/BSA, Molecular Probes) for 1 hr at room temperature. The cells were washed in PBS (twice), incubated for 5 min with TO-PRO-3 iodide (1:10000 in PBS) to stain the nuclei, washed again in PBS and mounted in Vectashield mounting medium (Vector Laboratories, Peterborough, UK). Cells were viewed with a Leica TCSNT confocal laser-scanning microscope.

Results

Caveolin is present in all PBMCs

None of the control antibodies gave any staining, as viewed by confocal microscopy (not shown). Similarly, the Alexa-conjugated antibodies did not stain cells in the absence of primary antibody. Caveolin-1, detected with the Santa Cruz polyclonal antibody could be seen in all cells present in PBMC cytospins by confocal microscopy (Fig. 1a-d). The protein was also detectable in magnetically sorted CD4⁺, CD8⁺, CD21⁺ and CD14⁺ cells (not shown). In addition, Western blotting demonstrated that the Santa Cruz polyclonal antibody against caveolin-1 bound a protein of the correct size (22000 MW) in monocytes, macrophages, DC, BAE-1 cells and human endothelial cells (Fig. 2a). In PBMCs, macrophages and DC caveolin-1 was detected only in the perinuclear region. In the endothelial cell line, BAE-1, staining for caveolin-1 was predominantly at the cell surface (Fig. 1e). The BD polyclonal antibody did not stain bovine PBMC, but did stain BAE-1 cells with a similar pattern to the Santa Cruz polyclonal antibody (Fig. 1e) and detected a 22000 MW protein in these cells by Western blotting (not shown).

Figure 1.

(a–d) Expression of caveolin-1 in bovine PBMC, detected by confocal microscopy. Cytospin preparations of PBMC were stained for caveolin-1 (red) and different cell differentiation markers (green; CD14 detected with anti-human CD14-labelled paramagnetic beads).(e–f) BAE-1 cells stained for caveolin 1 with two different antibodies; (e) Santa Cruz polyclonal antibody; (f) BD polyclonal antibody. Nuclei are stained blue with TO-PRO-3.

Figure 2.

Detection of (a) caveolin-1 and (b) caveolin-3 in bovine immune cells and human and bovine endothelial cells by Western blotting. Caveolin-1 was detected in all cells studied; HEC, human endothelial cell lysate; BAE-1, bovine aortic endothelial cells; MoDC, monocyte-derived dendritic cells; Mo, monocytes; Ma, macrophages; CD4, CD8 and CD21, sorted lymphocyte subsets. The Western blot shows strong labelling of a protein with a molecular mass of ∼22000. The polyclonal antibody against caveolin-3 detected proteins at ∼22000, 44000, 66000 and 210000 in HEC and ∼44000 and 66000 in bovine immune cells.

Caveolin-2 was not detected in any bovine cells by confocal microscopy or by Western blot (not shown). Antibody to caveolin-3 bound proteins of approximately 44000 and 66000 MW in all cells tested by Western blot (Fig. 2b), but could not be detected by confocal microscopy, except in FCK cells, in which the protein stained as small, discrete spots within the cytosol (not shown). In HEC lysate, Western blotting with the caveolin-3 polyclonal antibody produced four separate bands at approximately 22000, 44000, 66000 and 210000 MW; the latter gave the strongest signal (Fig. 2b).

Caveolin-1 co-localizes with Golgi and ERGIC in macrophages and DC

Using confocal microscopy, caveolin-1, as recognized by the Santa Cruz pAb, showed some co-localization with the Golgi protein golgin-97 and strong co-localization with the ERGIC, detected with anti-human ERGIC-53 mAb in monocyte-derived DC (Fig. 3a,c). Similarly strong co-localization between caveolin-1 and ERGIC was also observed in macrophages (not shown). Little or no co-localization was observed between caveolin-1 and the ER marker PDI (Fig. 3b)

Figure 3.

Confocal microscopy reveals co-localization between caveolin-1 (red) and Golgi and ERGIC, but not ER markers (green) in monocyte-derived DC. (a) Golgin-97; (b) PDI; (c) ERGIC-53. Overlap appears yellow; nuclei are stained blue with TO-PRO-3.

Discussion

Most of the research on caveolin in immune cells suggests that it is commonly found in myeloid, but not lymphoid cells. Caveolin and/or caveolae have been identified in human neutrophils,¹⁵ murine macrophages and mast cells^11,13 and human and bovine dendritic cells.¹⁴ However, with the exception of specific activated T-cell leukaemia cell lines¹⁰ or after transfection of human T-cell lines with caveolin 1,²⁴ neither caveolae nor caveolin have been found in T and B lymphocytes. In human and mouse T cells and mouse B cells, both caveolae and caveolin, at the protein and mRNA level, were found to be absent, despite the presence of glycolipid and glycosphingolipid-linked protein-enriched detergent-insoluble membrane microdomains normally associated with caveolae in other cells.¹⁷ Here we show that caveolin-1 is detectable by confocal microscopy and Western blotting in all bovine PBMCs, including CD4⁺, CD8⁺ and CD21⁺ lymphocytes and in CD14⁺ monocytes, as well as in cultured macrophages and monocyte-derived DC. Furthermore, the protein is concentrated mainly in the ERGIC of macrophages and DC. In preliminary investigations using confocal microscopy, we have found that both CD21⁺ and CD26⁺ human peripheral blood lymphocytes stain positively for caveolin-1 (unpublished data). Thus, the expression of caveolin by resting lymphocytes and APC reported here appears not to be restricted to cattle. Failure to detect caveolin/caveolae previously may reflect possible differences between lymphocyte cell lines and circulating lymphocytes, which might in turn imply a role for caveolin/caveolae in specific lymphocyte activation.

The absence of caveolin-1-staining on the cell membrane of these cells implies that they lack caveolae (particularly when compared with the staining in BAE-1 cells), while its presence in the ERGIC, and to a lesser extent the Golgi, may indicate a role in cholesterol transport and cell signalling. In human fibroblasts, caveolin cycles between the plasma membrane caveolae, ER and Golgi complex, then back to the caveolae. Movement from the ER to the Golgi is via the ERGIC and is microtubule-dependent.² Thus, in bovine immune cells, caveolin may, under normal conditions, be involved in the transport of fatty acids and cholesterol between the ER and Golgi. Although no co-localization was observed between caveolin-1 and the ER marker PDI, accumulation of caveolin in the ER may only be transient or dependent on specific conditions, such as cholesterol oxidation.² Similarly, expression of caveolin/caveolae at the cell surface may be transient or dependent on specific stimuli/activation.

Although it has been reported that different antibodies to caveolin may detect different pools of the protein – either on the cell membrane or in the Golgi area – depending on whether they are raised against the NH₂ or COOH terminus of the molecule, respectively,²⁵ the Santa Cruz polyclonal antibody (raised against the NH₂ terminus) stained BAE-1 cells at the cell membrane, ruling out the possibility that it might not detect caveolin at the cell surface of immune cells. The fact that the BD polyclonal antibody stained the cell membrane of BAE-1 cells, but did not stain bovine lymphocytes or APC would suggest that these two antibodies recognize different isotopes of the protein, possibly caveolin-1α and -1β,²⁶ one of which may not be present in bovine PBMC. The staining observed in bovine immune cells, particularly in macrophages and DC, is comparable to that seen in resident macrophages from the rat peritoneal cavity.¹¹ Elicitation of these macrophages (by injection of Freund's adjuvant) increases the expression of caveolin-1 and results in the formation of caveolae-like structures at the cell surface.¹¹ Similarly, in thioglycollate-elicited murine C3Heb/FJ peritoneal macrophages, caveolin-1 mRNA expression is up-regulated by treatment with lipopolysaccharide.¹² However, in a more recent study on resident and thioglycollate-elicited murine macrophages, caveolin-1 was found to be present at the cell surface, while caveolin-2 was found in the Golgi.²⁷ In the present study, caveolin-2 was not detected in bovine macrophages, DC, BAE-1, or FCK cells, though the antibody might not recognize the bovine protein. Caveolin-3 was detected in bovine immune cells by Western blotting, but not confocal microscopy. In this case, it is possible that the fixation method used for the confocal studies does not preserve or reveal the appropriate epitopes. Although caveolin-3 could be detected in FCK cells by confocal microscopy as small, discrete spots in the cytosol, this may reflect differences in the localization of the protein between epithelial and immune cells. The protein detected by Western blot is larger than the expected MW of 21000–24000,⁶ suggesting that it may be a complex of two or three caveolin molecules, or a storage form/precursor of caveolin-3.

Whether or not caveolae formation can be elicited in bovine immune cells or caveolin is involved in the response of these cells to immune challenge remains to be investigated, although the uptake and presentation of respiratory syncytial virus by bovine DC is sensitive to caveolae-disrupting agents.¹⁴ However, a lack of caveolae in immune cells, particularly in APC, could have important implications for the manner in which these cells interact with certain pathogens, including FimH-expressing Escherichia coli^13,18, which appear to target caveolae as a means of gaining safe entry into the cell and ensuring their subsequent intracellular survival.