Published online before print December 30, 2005
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
* Department of Rheumatology and Inflammation Research, University of Göteborg, Sweden;
Department of Cell and Molecular Biology, Lund University, Sweden; and
Hematology Section, Sahlgrenska University Hospital, Göteborg, Sweden
1 Correspondence: Department of Rheumatology and Inflammation Research, University of Göteborg, Guldhedsgatan 10, S 413 46 Göteborg, Sweden. E-mail: sara.pellme{at}rheuma.gu.se
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: chemokines • inflammation • subcellular organization • cell activation
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Neutrophils are equipped with different types of granules [16 17 18 ]: peroxidase-positive granules (primary or azurophil granules), which in many respects (but not all) resemble classical lysosomes [19 , 20 ], and two types of peroxidase-negative granules (secondary or specific granules and gelatinase granules) [21 , 22 ]. The granules contain a variety of different receptors and other membrane proteins, which are translocated to the cell surface as a consequence of granule fusion with the plasma membrane, concomitant to secretion of granule matrix proteins. In addition to the granules, neutrophils contain secretory vesicles. These are the most easily mobilized neutrophil organelles, formed through endocytosis at the endstage of differentiation in the bone marrow and thus, containing plasma proteins in their matrix [16 , 23 ]. Despite its origin as an endocytic organelle, the secretory vesicle has been suggested to be a CXCL-8-storing organelle [6 ]. Whether any of the organelles known so far can gain in protein content, e.g., de novo-synthesized cytokines/chemokines, also after release of the mature neutrophils from the bone marrow, is not yet known.
Here, the neutrophil subcellular localization of CXCL-8 has been studied thoroughly with regard to baseline levels and de novo-produced protein. We report, for the first time, that the de novo-synthesized CXCL-8 is stored in an organelle distinct from the four earlier, well-described granules/vesicles. Further, the CXCL-8 is retained in cytoplasts and is there partially cofractionating with the ER marker calnexin.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Quantification of CXCL-8
Production and release of CXCL-8 and the CXCL-8 content in intact neutrophils and cytoplasts as well as in gradient fractions were measured using a CXCL-8 enzyme-linked immunosorbent assay (ELISA; R&D Systems Europe, UK). Cells were resuspended in RPMI supplemented with heat-inactivated fetal calf serum (FCS; 10%) containing Pefabloc (1 mM) and Triton X-100 (1%), and cell-free supernatants were prepared by centrifugation of the cell suspension at 9000 g for 12 s, aspiration of the supernatant, and recentrifugation once prior to CXCL-8 analysis. Recombinant CXCL-8 was used to obtain a standard curve for quantification of the CXCL-8 amounts. The manufacturer’s limit of detection for CXCL-8 is 10 pg/ml.
Subcellular fractionation of neutrophils
Subcellular fractionation of neutrophils was performed at large according to Borregaard et al. [25
]. In short, neutrophils were treated with the serine protease inhibitor diisopropyl fluorophosphate (DFP; 6 µM) and ruptured by nitrogen cavitation (Parr Instruments Co., Moline, IL), and the postnuclear supernatant was fractionated in a Percoll gradient. Percoll solutions of different densities (see below) were layered from the bottom of the tube using a needle. The postnuclear supernatant was applied on top of the gradient, which was then centrifuged using a fixed-angle Beckman JA-20 rotor (37,000 g, 35 min, 4°C). Gradient fractions of 1.0 ml were collected, and localization of subcellular organelles in the gradients was determined by marker analysis of the fractions.
To isolate the main granule types from each other and from the plasma membrane and to separate the gelatinase granules from the specific granules, a modified, two-step gradient with Percoll solutions of 1.05 g/ml and 1.09 g/ml was used [26
]. This gradient separated the cell homogenate into four distinct fractions, the 
(azurophil granules), ß1 (specific granules), ß2 (gelatinase granules), and 
(plasma membrane/secretory vesicles) fractions.
The presence of the plasma membrane/secretory vesicle marker alkaline phosphatase (ALP), the specific granule marker vitamin B12-binding protein, and the azurophil granule marker myeloperoxidase (MPO) was determined as described in detail elsewhere [27 ]. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and Western blotting were used to detect the presence of the gelatinase and specific granule marker gelatinase, using a polyclonal rabbit anti-gelatinase antibody and a secondary peroxidase-conjugated goat anti-rabbit immunoglobulin G (IgG) antibody. The latency of ALP activity was used to identify the secretory vesicles [26 ]. Finally, samples from all fractions were assayed for content of CXCL-8 (see above).
To isolate secretory vesicles from the plasma membrane, a flotation gradient was used [19 , 28 , 29 ]. The postnuclear supernatant (7 ml) was mixed with a heavy Percoll solution (7 ml, 1.12 g/ml) and layered under a light Percoll solution (14 ml, 1.04 g/ml). Heavy Percoll solution (5 ml, 1.12 g/ml) was applied to the bottom of the tube. Relaxation buffer (5 ml) was applied on top of the gradient, which was centrifugated, and fractions were collected and analyzed as described above. Nonlatent ALP activity (measured in the absence of detergent) was used as marker for the plasma membrane, and as marker for the secretory vesicles, latent ALP activity (i.e., the activity measured in the presence of detergent minus the activity measured in the absence of detergent) was used.
Induction of CXCL-8 production and release
Neutrophils (107/ml) were resuspended in RPMI medium containing heat-inactivated FCS (10%) and incubated at 37°C in 5% CO2 with or without Escherichia coli LPS (10 µg/ml) for 4, 8, or 24 h, after which the amount of CXCL-8 was measured in the cells as well as in the cell culture medium. After 8 h of LPS treatment, the neutrophils had produced substantial amounts of CXCL-8 (150 ng/107 cells), of which approximately 40% had been released from the cells. The de novo synthesis continued for at least 16 h, and by 24 h, 60% was found in the cell-free supernatant. Also, in the absence of LPS, incubation of the cells over time induced production and secretion of CXCL-8. However, the amount was substantially lower (40 ng/107 cells after 8 h). Trypan blue exclusion revealed a cell integrity of >95% throughout the experiment, and the cells responded readily to phorbol 12-myristate 13-acetate (PMA) induction of superoxide anion production (data not shown) at all time-points, clearly indicating retained cellular functionality. The total endotoxin concentration of the FCS was less than 0.03 EU/ml (as stated by the manufacturer). All other reagents are considered LPS-free (less than 1 pg/ml), as determined by the Limulus assay.
Neutrophil nicotinamide adenine dinucleotide phosphate (NADPH)-oxidase activity
Cellular function was investigated by testing the cellular ability to activate the NADPH oxidase. The stimulus-induced production of superoxide anion was evaluated using an isoluminol-enhanced chemiluminescence (CL) system [30
] containing isoluminol (6x10–5 M) and horseradish peroxidase (HRP; 4 U). CL activity was measured in a six-channel Biolumat LB 9505 (Berthold Co., Wildbad, Germany), using disposable, 4 ml polypropylene tubes with a reaction mixture containing 106 neutrophils. The tubes were equilibrated in the apparatus for 5 min at 37°C, after which PMA (5x10–8 M) was added, and light emission measured continuously.
Electron microscopy (EM)
Samples for EM were prepared by pelleting 5 x 106 cells at 4°C immediately after addition of fixative (4% paraformaldehyde+0.1% glutaraldehyde). After incubation at room temperature for 1 h, the fixed pellets were dehydrated in ethanol and further processed for Lowicryl embedding [31
]. Sections were cut with a microtome and mounted onto nickle grids. For immunostaining, the grids were floated on top of drops of immune reagents displayed on a sheet of parafilm. Free aldehyde groups were blocked with 50 nM glycine, and the grids were then incubated with 5% (vol/vol) donkey serum in incubation buffer [0.2% bovine serum albumin-c in phosphate-buffered saline (PBS), pH 7.6] for 15 min. This blocking procedure was followed by overnight incubation with primary antibodies (dilution 1:50–1:100) at 4°C. After washing the grids in 200 mL incubation buffer, floating on drops containing the gold conjugate reagents (diluted 1:10–1:20 in incubation buffer) was performed for 60 min at room temperature. After further washes in incubation buffer, the sections were postfixed in 2% glutaraldehyde. Finally, sections were washed in destilled water and poststained with uranyl acetate and lead citrate and examined under the electron microscope.
The CXCL-8 antibody used in the EM experiments was tested in Western blots on whole cell neutrophil lysates and gave no staining other than an 8-kDa band (data not shown).
Preparation of cytoplasts
Cytoplasts were prepared according to Roos et al. [32
]. In short, neutrophils (108/ml) resuspended in 5 ml Ficoll solution (12.5% w/v) were layered on top of a prewarmed (2 h, 37°C), discontinuous density gradient (5 ml 17% w/v Ficoll solution on top of 5 ml 25% w/v Ficoll solution) in polycarbonate tubes. Cytochalasin B (cytb; 10 µg/ml) was present in all Ficoll solutions. The tubes were centrifugated (21,000 rpm, 30 min, 37°C) in a Beckman SW-27 swing-out rotor. After centrifugation, the cytoplasts were recovered in a band at the interface of the 12.5 and 17% Ficoll layers.
Subcellular fractionation of cytoplasts
Cytoplasts (approximately 108 cytoplasts in 5 ml) were fractionated on a modified flotation gradient as described above. In short, the cytoplasts were disrupted by nitrogen cavitation, mixed with a heavy Percoll solution (10 ml, 1.12 g/ml), and layered under a light Percoll solution (15 ml, 1.04 g/ml). Relaxation buffer (5 ml) was applied on top of the gradient. The gradient was centrifugated (37,000 g, 35 min), and 1.0 ml fractions were collected by aspiration from the bottom of the tube. Fraction content of ALP and CXCL-8 was determined using the same assays as described for the neutrophil fractionation. Annexin-I, which is present exclusively in the neutrophil cytosol [33
], was used as a marker for the cytosol, and calnexin was used as an ER marker [34
, 35
]. The distribution of annexin-I and calnexin in the gradient was determined by SDS-PAGE and immunoblotting of the fractions. Antibodies applied were for annexin-I, a monoclonal mouse anti-annexin-I antibody, followed by a peroxidase-conjugated secondary goat anti-mouse IgG antibody and for calnexin, a polyclonal rabbit anti-calnexin antibody, followed by a secondary anti-rabbit peroxidase-conjugated antibody.
Secretion of CXCL-8
Resting or LPS-treated (8 h, 10 µg/ml) neutrophils were stimulated with secretagogues inducing different levels of granule mobilization. For a mild stimulation, the cells were incubated at 37°C for 15 min in the presence of absence of formylmethionyl-leucyl-phenylalanine (fMLF; final concentration, 10–7 M). To induce more substantial degranulation, the cells were incubated in the presence of the cytoskeleton-disrupting agent cytb (5 min, 37°C; final concentration, 5 µg/ml) and were then stimulated with fMLF (15 min, 37°C; final concentration, 10–7 M) or ionomycin (10 min, 37°C; final concentration, 5x10–7 M). After stimulation, supernatants were collected by centrifugation (9000 g, 12 s), and the intact cells were resuspended in KRG (107 cells/ml) containing Pefabloc (a serine protease inhibitor; 1 mM) and Triton X-100 (1%), and the samples were analyzed for CXCL-8 content.
The mobilization of complement receptor 3 (CR3; CD18/CD11b) from the secretory vesicles, gelatinase, and specific granules to the cell surface was assessed in the cell populations by immunostaining and fluorescein-activated cell sorter (FACS) analysis. After fixation of the cells with ice-cold paraformaldehyde (PFA; 2%), the samples were washed twice in FACS wash (PBS, 0.02% NaN3, 10–4 M EDTA) and incubated at 4°C with phycoerythrin (PE)-conjugated anti-CD11b antibodies (10 µl/106 cells). Finally, the cells were washed twice and analyzed by FACScan.
Reagents and antibodies
Dextran, Ficoll-Paque, and Ficoll 70 were purchased from Pharmacia (Uppsala, Sweden). RPMI 1640 came from Göteborgs Termometerfabrik (Sweden) and FCS, from PAA Laboratories (Austria). Triton X-100, fMLF, isoluminol, adenosine 5'-triphosphate, EGTA, DFP, LPS (E. coli serotype O111:B4; L-2630), and cytb were obtained from Sigma Chemical Co. (St. Louis, MO). Pefabloc was from Boehringer-Mannheim (Germany) and PE-conjugated anti-CD11b antibodies, from BD Biosciences (Mountain View, CA). The CXCL-8 ELISA was from R&D Systems Europe (UK), ionomycin was from Calbiochem (La Jolla, CA), and Percoll was from Amersham Bioscience Europe (Germany). The SDS was purchased from Fluka Chemie (Buchs, Switzerland). Antibodies against lysosome-associated membrane protein-1 (LAMP-1) used in EM were a kind gift from Dr. Sven Carlsson (Umeå University, Sweden). HRP-labeled secondary antibodies came from Dako (Denmark), as did the rabbit -human MPO antibody used in EM. The rabbit -human gelatinase [matrix metalloproteinase-9 (MMP-9)] antibodies came from Chemicon (Temecula, CA). Monoclonal anti-annexin-I antibody was from Transduction Laboratories (Mamhead, UK). The human CXCL-8 monoclonal antibody used in EM was purchased from Accurate Chemicals (New York, NY). The polyclonal anti-calnexin antibody was a kind gift from Professor Ari Helenius (Institute of Biochemistry, ETH-Hoenggerberg, Zurich, Switzerland).
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
3 ng/107 cells) in resting peripheral blood neutrophils, corresponding well to previously published results [9
, 36
, 37
].
To study the subcellular localization of CXCL-8, we performed Percoll gradient experiments with resting neutrophils. The subcellular localization of CXCL-8 was first determined by fractionation of the cells on a two-layer discontinuous Percoll gradient. In such a gradient, four known subcellular organelles can be identified easily: the azurophil granules (the band, measured as MPO content), the specific granules (the ß1band, measured as vitamin B12-binding protein and gelatinase content), the gelatinase granules (the ß2band, measured as gelatinase, but no vitamin B12-binding protein content), and the plasma membrane/secretory vesicles (the band, measured as total ALP activity). The gradient presented here (Fig. 1A
) follows this pattern of marker distribution, apart from a slight contamination of specific granules in the gelatinase granule band, seen as a minor content of vitamin B12-binding protein. We found the CXCL-8 content in the light membrane fraction (the band) enriched for secretory vesicles and plasma membranes. This is in line with the report by Kuhns and Gallin [6
], showing that CXCL-8 in resting neutrophils is localized in a light membrane fraction, which they claim to be the secretory vesicles.
 View larger version (26K): [in a new window] |
Figure 1. The CXCL-8 in resting neutrophils is localized in a light membrane fraction distinct from the secretory vesicles. Resting neutrophils were fractionated in a two-step Percoll gradient (A). Localization of neutrophil organelles in the gradient was shown by marker analysis of the fractions [arbitrary units (AU)]: MPO (marker for the azurophil granules,  ); vitamin B12-binding protein (marker for the specific granules,  ), and total alkaline phosphatase activity (marker for the plasma membrane/secretory vesicles, •). Presence of gelatinase is shown by immunoblotting. The cytosol was localized to Fraction 27 and above (data not shown). The presence of CXCL-8 in the gradient was determined by ELISA ( ). Resting neutrophils were further fractionated in a flotation gradient (B). Localization of the azurophil () and specific/gelatinase granules (ß) was determined by the presence of MPO and vitamin B12-binding protein, respectively, and the plasma membrane and secretory vesicles were localized by nonlatent alkaline phospahatase (•) and latent alkaline phosphatase ( ), respectively. The cytosol was localized to Fraction 29 and above (data not shown). The presence of CXCL-8 in the gradient was shown by ELISA ( ).
|
1, latent ALP activity) from the plasma membranes (2, nonlatent ALP activity). In contrast to the conclusion drawn by Kuhns and Gallin [6
], CXCL-8 did not colocalize with the secretory vesicles but was instead recovered in a broad peak between the secretory vesicles and the plasma membrane peaks. In some experiments, a minor part of the CXCL-8 was recovered in the ß-fraction, of which 70% colocalized with the gelatinase granules (not shown by figure). The amount of CXCL-8 in resting neutrophils was too low to achieve reliable staining for immunoelectron microscopy (data not shown) and confirming morphological experiments, could thus not be performed. Therefore, we decided to study neutrophils that had been induced to produce/contain increased amounts of CXCL-8 by bacterial LPS. LPS is a strong inducer of transcription of cytokine genes in macrophages and other immune-competent cells and previous studies have shown that also neutrophils respond to LPS by production of cytokines [12 13 14 15 ].
Immunoelectron microscopy reveals CXCL-8 in an organelle distinct from the classical granules
The classical neutrophil organelles are formed during a limited period of cell development. They are then considered to be storage pools of proteins and membrane, remaining static until mobilized for use during the inflammatory process. As CXCL-8 was synthesized de novo and stored in neutrophils in increasing amounts, we wanted to determine if any of the earlier defined organelles are used for this purpose or if this newly synthesized CXCL-8 is stored elsewhere.
EM has previously been used to visualize neutrophil granules and their contents [18 , 20 ]. When using immunogold labeling of thin sections to investigate CXCL-8 distribution in LPS-stimulated cells, the staining was detected inside morphologically distinct organelles (Fig. 2A ). Further, double immunostaining for CXCL-8 and markers for defined neutrophil granules, i.e., MPO (azurophil granules) and gelatinase (specific and gelatinase granules), revealed micrographs, which clearly showed the presence of CXCL-8 in vesicles separate from organelles labeled with the other markers (Fig. 2B and 2C) . The MPO-containing azurophil granules were larger than the CXCL-8-containing organelles (Fig. 2B) and appeared less electron-dense. The gelatinase and specific granules were smaller than the azurophil granules and of similar size to the CXCL-8-containing organelles but more elongated and irregular in shape (Fig. 2C) .
 View larger version (167K): [in a new window] |
Figure 2. Newly produced CXCL-8 is localized in distinct subcellular organelles and does not colocalize with markers for the classical granules. The subcellular localization of de novo-synthesized CXCL-8 was examined by EM. CXCL-8 was visualized by binding of anti-CXCL-8 antibody followed by detection with immunogold-labeled secondary antibody in thin sections of LPS-activated (10 µg/ml, 8 h, 37°C) neutrophils. The LPS-induced CXCL-8 (gold particles) was found inside organelles, which appeared round or somewhat elongated and of medium size as compared with other morphologically distinct organelles (A). Double immunolabeling with markers for defined neutrophil granules was performed: (B) CXCL-8 (small gold particles) and MPO (marker for azurophil granules, large gold particles); (C) CXCL-8 (small) and gelatinase (marker for specific and gelatinase granules, large); and (D) CXCL-8 (large) and LAMP [marker for multivesicular bodies (MVB) and multilaminar compartments (MLC), small].
|
CXCL-8 is localized in a light membrane fraction in LPS-treated neutrophils
The EM results were confirmed by biochemical analysis. Subcellular localization of CXCL-8 in LPS-treated peripheral blood neutrophils was determined by fractionation of the cells on a two-layer, discontinuous Percoll gradient. Marker analysis of the fractions revealed four subcellular fractions (Fig. 3
): i.e., the azurophil granules (the band), the specific granules (the ß1 band), the gelatinase granules (the ß2band), and the plasma membrane (the band, measured as nonlatent ALP activity). No secretory vesicles (measured as latent ALP activity) could be detected, indicating that the secretory vesicles had been mobilized completely during the 8-h incubation with LPS.
 View larger version (23K): [in a new window] |
Figure 3. De novo-synthesized CXCL-8 is localized to a light membrane fraction in LPS-treated neutrophils. Activated neutrophils (LPS, 10 µg/ml, 8 h, 37°C, 5% CO2) were fractionated in a two-step Percoll gradient. Localization of neutrophil organelles in the gradient was shown by marker analysis: nonlatent alkaline phosphatase activity (in the absence of detergent, marker for the plasma membrane, •); latent alkaline phosphatase activity (total minus nonlatent, marker for the secretory vesicles, ); MPO activity (marker for azurophil granules, ); vitamin B12-binding protein (marker for specific granules, ); and gelatinase (marker for gelatinase granules, determined by densitometry of Western blots,  ). The presence of CXCL-8 in the gradient was determined by ELISA (). The amount of markers is given in arbitrary units. The cytosol is localized to fractions 27 and above (data not shown).
|
band) enriched for plasma membranes. Hence, CXCL-8 is not cytosolic but is localized in a light membrane organelle, cofractionating with the plasma membrane, but most important, different from the (nonpresent) secretory vesicle. Comparing with the distribution in resting neutrophils, CXCL-8 appears roughly at the same place in the gradient, suggesting that the CXCL-8-containing organelle is the same or similar in the two cell types. These data, together with the immunoelectron microscopy results, convincingly show that CXCL-8, which has been synthesized upon LPS treatment, is localized in an organelle different from all the known granules and vesicles in human neutrophils.
Neutrophil cytoplasts retain CXCL-8
To further investigate the features of the CXCL-8-storing organelle, we prepared so-called cytoplasts [32
]. During cytoplast preparation, the nucleus and granules exit the cell, leaving behind the plasma membrane surrounding the cytosol. We wanted to examine if the CXCL-8-storing organelle leaves the cytoplasts together with the granules or if it behaves differently from the classical organelles.
Intact, resting neutrophils and cytoplasts were assayed for CXCL-8 content and were found to contain CXCL-8 of roughly the same amounts (2.6±1.5 ng/107 neutrophils and 4.8±1.8 ng/107 cytoplasts; Fig. 4A ), suggesting that the CXCL-8 organelle is retained in the cytoplasts. This was confirmed by subcellular fractionation. As the CXCL-8 organelle was partly separated from the plasma membrane in intact neutrophils by the use of a flotation gradient (see above), we used a modification of this technique to fractionate the cytoplasts.
 View larger version (12K): [in a new window] |
Figure 4. Cytoplasts contain CXCL-8 of amounts similar to resting neutrophils, and the CXCL-8 organelle can be isolated by Percoll fractionation. Cytoplasts were prepared by centrifugation of neutrophils in a Ficoll gradient containing cytb. (A) The amount of CXCL-8 in the cytoplasts (solid bar) was comparable with that of intact neutrophils (open bar). Data are given as mean + SD, n = 3. n.s, Not significant. (B) Cytoplasts were fractionated in a flotation gradient, and marker analysis of the fractions was performed: The plasma membrane is seen as alkaline phosphatase activity (•), the cytosol is represented by annexin-I (shown by immunoblot, middle panel), and the ER is represented by calnexin (shown by immunoblot, bottom panel). The presence of CXCL-8 in the gradient was determined by ELISA (). The amount of marker is given as arbitrary units.
|
CXCL-8 partly colocalizes with the ER marker calnexin
As compartments of the constitutive secretory pathway have low isopycnic density, such structures might be retained in the granule-depleted cytoplasts and possibly contain CXCL-8. In accordance with this, membranes containing the ER marker calnexin were indeed present in the cytoplasts from resting neutrophils, and the distribution of this marker overlapped that of IL-8 (Fig. 4B
, bottom panel). This was only to be expected, as ER/Golgi structures (identified by the Golgi marker galactosyl transferase) have previously been found in the light fractions of Percoll gradients [38
]. However, calnexin and CXCL-8 in the cytoplast gradient only partly colocalized; i.e., the peak value for CXCL-8 was found in Fraction 13, and the calnexin peak was detected in Fraction 15. Thus, it is reasonable to assume that only a portion of the ER-derived structures also contains CXCL-8, if at all the CXCL-8 is localized in these structures.
Mobilization of the CXCL-8-containing organelle in response to secretagogues differs between resting and LPS-treated neutrophils
As some of the neutrophil organelles (secretory vesicles, gelatinase granules, and specific granules) are more sensitive to stimulated exocytosis than others (i.e., azurophil granules and MVB/MLC), we wanted to find out whether mobilization of the CXCL-8-containing organelle to the plasma membrane could be induced by the addition of known secretagogues.
When resting neutrophils were stimulated with the chemoattractant fMLF alone or with cytb/ionomycin (strong secretion protocol), the marker molecule CR3 (CD18/CD11b; localized to the secretory vesicles and specific/gelatinase granules) was mobilized to the cell surface (Fig. 5A ). The secretagogues also induced a rapid secretion of CXCL-8 (Fig. 5C) , and the amount of secreted protein correlated with the strength of the secretion protocol. These data strongly suggest that mobilization of the CXCL-8-containing organelle in resting neutrophils is inducible by secretagogues.
 View larger version (25K): [in a new window] |
Figure 5. The CXCL-8 organelle in resting and LPS-treated neutrophils responds differently to secretagogues. Resting neutrophils (A and C) or neutrophils incubated in the presence of LPS (10 µg/ml) for 8 h (B and D) were used. The cells were kept on ice (control) or further stimulated at 37°C with fMLF (10–7M, 15 min) or cytb (5 µg/ml, 5 min) followed by ionomycin (5x10–7 M, 10 min). After centrifugation, the supernatants were analyzed for secreted CXCL-8 by ELISA, and results are presented in ng/107 neutrophils. The intact cells were fixed (PFA, 2%), incubated with PE-conjugated CD11b (subunit of CR3) antibody, and analyzed by FACScan. The results are given as percent of CR3 expression in control cells. Mean + SD of three or four independent experiments is shown. Student’s t-test (two-tailed, paired) was used to determined statistical significance. *, P < 0.05.
|
Hence, although the CXCL-8 organelles fractionate similarly in resting and LPS-treated cells, they do not share the ability to respond to secretagogues. Whether this is a result of the fact that the organelles are not identical or whether the CXCL-8 organelle has lost its responsiveness during the LPS treatment can, at this time, only be speculated upon.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
In thin sections of LPS-stimulated neutrophils, a vast number of granules of different shapes and electron density can be seen. Among these are the azurophil granules, originally described as lysosome-like, peroxidase-positive granules, and the specific/gelatinase granules, originally said to comprise all peroxidase-negative organelles but now, more specifically, defined as granules containing lactoferrin, vitamin B12-binding protein, and/or gelatinase. Specific/gelatinase granules are somewhat smaller than azurophil granules and more irregular in shape. The smallest known neutrophil secretory organelle is the secretory vesicle, having an elongated shape and containing plasma proteins and plasma membrane components [16 , 39 ]. From double-labeled immunoelectron micrographs shown in this study, we conclude that the CXCL-8 organelle is different from all the classical granules, shown by different localization of the immunostaining of CXCL-8 and other granule markers. Further, CXCL-8 was not localized in the recently described MLC and MVB, which contain the lysosomal markers LAMP-1 and -2 [20 ]. These organelles differ from the granules with regard to morphology and have been suggested to be the remains of a neutrophil Golgi apparatus as a result of their content of mannose-6 phosphate. Such organelles would have been a possible location for the de novo-synthesized CXCL-8, but no colocalization of CXCL-8 and LAMP was shown by immunoelectron microscopy.
The results from immunoelectron microscopy were confirmed by subcellular fractionation. The localization of CXCL-8 in Percoll gradients of resting and LPS-treated cells indicated that the CXCL-8-containing organelle(s) is of low density, colocalizing at least in part with the plasma membrane and the secretory vesicles (in resting cells). The azurophil, specific, and gelatinase granules were all localized much further down in the gradient, i.e., considerably denser, confirming that CXCL-8 is not localized to these organelles. Regarding the secretory vesicles, they are present only in the resting cells, as LPS treatment induces degranulation of these easily mobilized organelles (ref. [40 ] and data not shown). When separating the secretory vesicles from the plasma membranes by the use of a flotation gradient, the distribution profile of latent ALP (secretory vesicle marker) was different from that of CXCL-8, showing that the organelles storing these two proteins are not identical.
The study by Kuhns and Gallin [6 ] shows that CXCL-8 in resting neutrophils is localized in a light membrane fraction, which they claim to be the secretory vesicles. Their subcellular localization is in line with what we find when fractionating resting cells on a two-layer sedimentation gradient, where CXCL-8 colocalizes with latent and nonlatent alkaline phosphatase, the marker for secretory vesicles and plasma membrane, respectively. However, we have previously found that separation of the light membranes, i.e., the plasma membrane and the secretory vesicles (which are endocytic in origin and therefore, similar to the plasma membrane in composition), cannot be achieved by sedimentation centrifugation. Instead, we use flotation to separate these organelles. Using a flotation gradient, we find the CXCL-8 peak to be positioned separate from the plasma membrane and the secretory vesicles. We therefore conclude that the CXCL-8-containing organelle is not identical to the secretory vesicle but that it is a light organelle of similar density. Further supporting the conlusion that CXCL-8 is not stored in the secretory vesicle is the fact that "cells" devoid of secretory vesicles, i.e., HL-60 cells (data not shown) and cytoplasts, contain CXCL-8 of similar amounts as intact neutrophils. The cytoplast content of CXCL-8 flotated in the Percoll gradient, suggesting that the compartment storing the cytokine is retained in the cytoplast, while none of the classical granules and vesicles are. This finding gives support to our claim that CXCL-8 is neither stored in any of the classical granules nor in the secretory vesicles.
As a result of the low isopycnic density of compartments involved in the constitutive secretory pathway, although not described previously in mature neutrophils [41 ], it is reasonable to believe that such compartments would also be retained in cytoplasts. When examining the localization of the ER protein calnexin in cytoplasts as well as in LPS-treated cells, we found a partly overlapping distribution profile between this protein and CXCL-8. This suggests a possible relation between CXCL-8 storage and the constitutive secretory pathway in neutrophils, which is intriguing and will be subject to further investigation. However, the nonidentical distribution profiles indicate that CXCL-8 is localized in a subpart of the ER, or it is localized in a totally different organelle, cofractionating with the ER. As the presence of ER/Golgi content in mature neutrophils has neither been explored nor discussed in detail in the literature so far, it is evident that this is an area of research that needs to be explored.
The different composition of membrane proteins and intravesicular content between the neutrophil granules and vesicles has been suggested to reflect differences in the time of biosynthesis of these granules/vesicles in the maturing neutrophil (see refs. [17
, 42
] for a description of the sorting-by-timing hypothesis). The time of synthesis of granule/vesicle proteins and formation of the organelles appear to define not only the localization of proteins to a certain granule type but also to determine the sequence in which the granules/vesicles are used during the extravasation and microbial elimination processes. Hence, in the mature neutrophil, mobilization of granules and vesicles to the plasma membrane seems to occur in the reverse order of their formation, the secretory vesicles being the most easily mobilized organelle, followed by the gelatinase granules and the specific granules. Whether the CXCL-8 organelle fits into this pattern of synthesis and mobilization can at this time only be speculated upon. If CXCL-8 is to be used early in the inflammatory process, it should, according to the model of formed-first-released-last, be stored in an easily mobilized vesicle formed late in the differentiation process. Indeed, the CXCL-8 present in resting peripheral blood neutrophils is stored in an easily mobilized organelle (but when this organelle is formed remains to be determined). Conversely, CXCL-8 can be involved at a later phase of the immune response, e.g., in attracting a second wave of neutrophils to an inflammatory site, where the present cells are primed by bacterial and host components. Further, the late secretion of CXCL-8 may desensitize the cells to additional chemotactic stimulation of the same chemokine, resulting in arrest of the cells in the specific area. In such a scenario, CXCL-8 will be produced in large amounts (in direct or indirect response to, e.g., LPS stimulation or cytokine production, e.g., endogenously produced tumor necrosis factor [12
, 43
]) and constitutively secreted—a secretion we have shown not to be influenced by secretagogues. Such production and release are, however, not in line with the stipulated hypotheses of protein synthesis and mobilization in neutrophils mentioned above.
It might of course be that part of the neutrophil content of CXCL-8 originates from an endocytotic process of released material. It should, however, be noticed that the effective concentration of 50% (EC50) values for CXCR1- and CXCR2-dependent CXCL-8 uptake are much higher (50 and 20 nM, respectively [44 ]) than the amount of CXCL-8 present in blood of normal individuals (approximately 10 pM) [45 ]. This suggests that the small amounts of CXCL-8 in resting neutrophils are produced and sorted by the cells themselves. Although the CXCL-8 produced by LPS-treated cells is much higher than in resting cells, the amount of CXCL-8 released from the cell (which after 24 h reaches nM levels) is still below the EC50 for uptake. The neutrophil content of CXCL-8 at the time-point chosen for our localization/immunogold-labeling experiments (8 h of LPS stimulation) is even lower and thus, most probably endogenously produced.
An intracellular compartment, cofractionating with light membranes (plasma membrane/secretory vesicles) in density gradients and differing from the defined granule populations on the basis of overall collective features, such as marker content, size, and shape, was described recently [46 ]. The organelle contains the tissue inhibitor of MMP-1 (TIMP-1) but is devoid of markers for the secretory vesicle, making it a possible candidate for the CXCL-8-storing organelle. However, the TIMP-1 and CXCL-8-containing organelles differ in at least one aspect: The cytoplasts contain small amounts of TIMP-1 (data not shown), and the content of CXCL-8 is in the same range as for neutrophils, suggesting that the TIMP-1-containing organelle leaves the neutrophil during cytoplast preparation, while the CXCL-8 organelle stays intact.
Altogether, the role of the CXCL-8 pool, increasing with time, is obscure, but the fact that LPS as well as several neutrophil-stimulating agents and the process of extravasation induce a massive production of CXCL-8 indicates that this is a significant part of the neutrophil role in the inflammatory response.
Received May 9, 2005; revised November 4, 2005; accepted November 7, 2005.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, and IL-1 ß J. Cell. Physiol. 154,478-485[CrossRef][Medline], PMA, and LPS on plasma and cell-associated IL-8 in human leukocytes Thromb. Res. 113,75-83[CrossRef][Medline]This article has been cited by other articles:
 |
|
 |
|
  M. P. Simons, K. G. Leidal, W. M. Nauseef, and T. S. Griffith TNF-related apoptosis-inducing ligand (TRAIL) is expressed throughout myeloid development, resulting in a broad distribution among neutrophil granules J. Leukoc. Biol., March 1, 2008; 83(3): 621 - 629. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Reutershan, R. E. Cagnina, D. Chang, J. Linden, and K. Ley Therapeutic Anti-Inflammatory Effects of Myeloid Cell Adenosine Receptor A2a Stimulation in Lipopolysaccharide-Induced Lung Injury J. Immunol., July 15, 2007; 179(2): 1254 - 1263. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
