Introduction

Top Abstract Introduction Methods Results Discussion References

The LTs are biologically active lipids derived from the metabolism of arachidonic acid in response to a number of immunologic and inflammatory stimuli (Samuelsson, 1983). Arachidonic acid is converted, by the action of 5-lipoxygenase, to a hydroperoxy intermediate, which is then dehydrated to form LTA₄. In turn, LTA₄ is either converted to LTB₄ via the action of LTA₄ hydrolase or is conjugated with reduced glutathione by LTC₄ synthase to form the sulfidopeptide leukotriene, LTC₄. Leukotriene C₄ may then be converted to LTD₄ by -glutamyltranspeptidase and, LTD₄ may be subsequently metabolized to LTE₄ by cysteinylglycine dipeptidase (Samuelsson et al., 1987). The sulfidopeptide LTs (LTC₄/D₄/E₄) are potent vaso- and broncho-constrictors (Dahlén et al., 1980; Drazen et al., 1980; Greenwald et al., 1984). In addition, the LTs increase permeability in several microvascular beds including, guinea pig skin (Williams and Piper, 1980) and hamster cheek pouch (Dahlen et al., 1981; Björk et al., 1983). Moreover, in animal models of acute lung injury (neutrophil dependent or independent), increased levels of LTs occur in the edema fluid pari passu with enhanced microvascular permeability (Lonigro et al., 1989). However, the cellular source of LTs for interaction with the microvasculature remains unknown.

Pericytes can be thought of as `the smooth muscle cells' of the capillaries and postcapillary venules. As with smooth muscle cells, pericytes are highly contractile cells (Kelley et al., 1987) which, in vitro, express the smooth muscle type -isoform of actin (Herman and D'Amore, 1985). They form a discontinuous envelope around, and share a common basement membrane with, the underlying endothelial cells. Pericytes extend processes along the long axis of, as well as around, the capillary and these long processes can make contact with several endothelial cells (Shepro and Morel, 1993; Sims, 1991). Moreover, pericytes produce several bioactive autacoids; e.g., PGI₂, PGE₂, PGF₂ and thromboxane A₂ (Hudes et al., 1988; Eskenasy and Tasca, 1988). By virtue of their capacity to synthesize and/or respond to inflammatory mediators (Dodge et al., 1991), pericytes have been suggested to participate in the regulation of capillary permeability (Miller and Sims, 1986; Sims et al., 1990; Murphy and Wagner, 1994).

We propose that pericytes, when supplied with the appropriate substrate, synthesize sulfidopeptide LTs, compounds that have been linked to increased microvascular permeability. Thus, this study investigates the ability of pericytes to contribute to vascular LT synthesis, either directly, or by metabolizing LTA₄ supplied from a potentially important source such as, the PMN.

	Methods

Top Abstract Introduction Methods Results Discussion References

Cell culture. BRPs were obtained using a modification of the technique of Gitlin and D'Amore (1983). Cow eyes from the local abattoir were transported, in ice-cold saline, to the laboratory. Eyes were soaked in betadine for 10 min then rinsed in PBS (10 min), containing (in mM) NaCl 140, Na₂HPO₄ 10, KH₂PO₄ 1.5 and KCl₃. The eyes were bisected approximately 5 mm posterior to the limbus and the retinas were removed and placed in PBS. Retinas were rinsed three times in sterile PBS and pigmented epithelium was removed. The retinas were minced thoroughly and centrifuged (800 × g for 8 min), before digestion with .2% collagenase (type II; 30 min at 37°C) supplemented with penicillin/streptomycin (3%) and amphotericin B (1%). Thereafter, the digest was passed over a filter (Nitex, 100 µm). The filtrate was centrifuged and the pellet was washed three times in MEM with 10% fetal bovine serum, 1% penicillin/streptomycin and 1% amphotericin B. Two bovine retinas were used to seed one T75 flask (Falcon, Bedford, MA). The BRPs were cultured in MEM with 10% fetal bovine serum, 1% penicillin/streptomycin and 1% amphotericin B and maintained at 37°C in a 5% CO₂ atmosphere. After 24 hr the medium was changed to remove unattached cells.

Cell identification. BRPs were identified based on the morphology of the cells and by immunofluorescence staining, as previously reported (Lonigro et al., 1996). The pericyte in culture is a large, flat, stellate cell with long, slender processes, membrane ruffling and short, broad filopods (Shepro and Morel, 1993). Although these characteristics are typical of pericytes, immunofluorescence staining was also performed to document that all cultures contained exclusively this cell type. Cells were grown on glass coverslips, and were immunostained for alpha-smooth muscle actin, 3G5 antigen, a specific marker for pericytes within the microvasculature (Nayak et al., 1988) and for acetylated low-density lipoprotein. Cells, which were to be stained for alpha-smooth muscle actin or 3G5, were rinsed with PBS and fixed with 4% formalin. Cells to be stained for alpha-smooth muscle actin were then permeabilized with acetone. The primary antibody was either anti-alpha-smooth muscle actin (1:400; Sigma Chemical Co., St. Louis, MO) or anti-3G5 ganglioside (1:100). The secondary antibody was fluorescine-labeled goat anti-mouse IgG (1:64 and 1:20, for alpha-smooth muscle actin or 3G5, respectively; Sigma). Primary and secondary antibodies were incubated sequentially with the cells for 30 min each at 37°C. For identification of acetylated low density lipoprotein, cells were incubated (4 hr at 37°C) with the DiI-acetylated low density lipoprotein, rinsed with PBS, and fixed with neutral, buffered 4% formalin. Washed and dried coverslips were mounted on slides using Aqua-Mount and examined under a Nikon Diaphot fluorescent microscope.

Preparation of human neutrophils. Human neutrophils were prepared according to the method of Burkey and Webster (1993). Heparinized (10 U/ml) blood was obtained from consenting, healthy volunteers by venipuncture. Erythrocytes were separated from leukocytes by sedimentation (1 × g, 22°C, 30 min) in a 2:3:5 mixture of 6% Dextran T-500 in PBS (120 mM NaCl, 6 mM Na₂HPO₄, 3 mM KH₂PO₄, pH 7.2-7.4) and blood. Suspended leukocytes were removed, centrifuged (22 × g, 22°C, 10 min) and washed in PBS. Leukocytes were then subjected to density gradient centrifugation (300 × g) in Lymphoprep for 30 min at 22°C. The cell pellet was washed once with PBS and contaminating erythrocytes were removed by hypotonic lysis. Cell preparations were 90 ± 1.7% PMNs, 3.3 ± 0.7% eosinophils, 0.1 ± 0.4% basophils, 3.6 ± 1.1% lymphocytes, 0.7 ± 0.5% monocytes and 1.8 ± 0.6% macrophages (n = 12). Cells were >97% viable as determined by trypan blue exclusion.

Preparation of bovine neutrophils. At the slaughterhouse, bovine blood was collected directly into a flask containing heparin (final concentration, 10 U/ml). Erythrocytes were separated from leukocytes by sedimentation of the whole blood (1 × g, 22°C, 30 min) in a 2:3:5 mixture of Dextran T-500 in PBS (120 mM NaCl, 6 mM Na₂HPO₄, 3 mM KH₂PO₄, pH 7.2-7.4). Suspended leukocytes were removed, centrifuged (22 × g, 22°C, 10 min) and washed with PBS. Leukocytes were then subjected to density gradient centrifugation (300 × g, 22°C). The cell pellet was washed once with PBS and contaminating erythrocytes were removed by hypotonic lysis. These cell preparations routinely result in >90% neutrophils, which are >97% viable as determined by trypan blue exclusion.

Protein measurement. Cell protein content was measured after the removal of the incubation supernatant, which contained medium and, in the case of the coculture experiments, the majority of the neutrophils. The pericytes were washed several times before protein content was measured. Cell protein was measured using the BCA protein assay (Pierce, Rockford, IL) which was adapted for use with microtiter plates. Cell pellets were hydrolyzed with NaOH (1 M) overnight and diluted to .1 M before conducting the protein assay. Samples (10 µl) were incubated with 200 µl BCA reagent for 2 h and the absorbance was measured at 550 nm in a plate reader (model EL-311 multiwell plate reader, Bio-Tek). Sample protein concentration was based on bovine serum albumin (BSA) standards.

EIA. Concentrations of LTC₄/D₄/E₄ were determined using an enzyme-linked immunoassay (EIA) kit (Amersham), with antiserum directed against the sulfidopeptide LTs. The enzymatic tracer was LTC conjugated to peroxidase. The sulfidopeptide LT antiserum exhibits the following cross-reactivities, calculated at 50% bound (B/B_o): LTC₄ and LTD₄ 100%, LTE₄ 70% and LTB₄ 0.3%. The cross-reactivity for other eicosanoids was < .006%. Absorbance at 450 nm was measured in a plate reader (model EL-311 multiwell plate reader, Bio-Tek, Winooski, VT). The sensitivity of the assay is 15 pg/ml. The assay was performed on 50 µl of the medium. The concentration of LTC₄/D₄/E₄ was determined by comparison to a log-logit transformation of the standard curve to LTC₄ included on each plate. In the coculture experiments, mean LT production by PMNs alone was subtracted from the amount of LTs produced by coincubation of pericytes and PMNs, thereby allowing results to be expressed as picograms LT per µg pericyte protein.

Northern blot analysis. Total RNA was isolated from 1 to 5 × 10⁹ cells with Trizol reagent (Gibco, Grand Island, NY) as described by the manufacturer. Approximately, 25 µg of total RNA were electrophoresed on 0.8% agarose, 1.1 M formaldehyde gel and transferred to Hybond N^+-nylon membranes (Amersham Corp., Arlington Heights, IL). The membranes were baked for 2 hr at 80°C under vacuum. A human 5-lipoxygenase specific probe was labeled by random priming (Rediprime, Amersham) with [-³²P]CTP (Amersham, 600 Ci/mmol) to a specific activity of 10⁹ cpm/µg DNA. The membranes were hybridized with the probe for 2 hr at 65°C, washed twice at 42°C with 5 × SSPE (20 mM EDTA, 1 M NaCl, 50 mM NaH₂PO₄) plus 0.1% SDS and once at 65°C with 0.1 SSPE plus 0.1% SDS. Quantification was done either by autoradiography at -20°C or by phosphorimage analysis (Molecular Dynamic Corp., Sunnyvale, CA).

Materials. Trypsin/EDTA, type II collagenase, penicillin/streptomycin, amphotericin B, DMSO, esculetin, alpha smooth muscle actin antibody, fluorescine-labeled goat anti-mouse IgG, arachidonic acid and all buffer salts were purchased from Sigma. Minimal essential medium was obtained from Gibco and fetal bovine serum from Washington University Tissue Culture Support Center (St. Louis, MO). Formalin was obtained from Mallinckrodt (St. Louis, MO), DiI-acetylated low density lipoprotein from Biomedical Technologies Inc. (Stoughton, MA), Dextran T-500 from Pharmacia Fine Chemicals (Uppsala, Sweden), Lymphoprep from Accurate Chemical and Scientific Corp. (Westbury, NY) and Aqua-Mount from Lerner Laboratories (Pittsburgh, PA). The LTC₄/D₄/E₄ EIA kit was supplied by Amersham. The 3G5 antibody was a generous gift from Dr. Ivan O. Haefliger, Bascom Palmer Eye Institute, Miami, FL.

Experimental protocols. Three experimental groups were studied, namely: (i) BRPs alone (n = 4), (ii) PMNs alone (n = 9-15) or (iii) BRPs with PMNs (n = 9-15). In group (iii), the PMNs (1 × 10⁶) were added to the BRPs 30 min prior to the start of the experiment. All experimental groups were then treated with either the lipoxygenase inhibitor, esculetin (1 × 10⁴ M) or with its vehicle (0.1 M Na₂CO₃) for 30 min at 37°C. This was followed by incubation with the calcium ionophore A23187 (0-1 µM) for 30 min.

Statistical analysis. All values in the figures and text are expressed as mean ± S.E.M. of n observations. Statistical evaluation of the data was by Student's t test for unpaired determinations or by analysis of variance followed by the Bonferroni post test. P < .05 was considered significant.

	Results

Top Abstract Introduction Methods Results Discussion References

BRPs alone produced undetectable amounts of sufidopeptide LTs, either basally, or after stimulation with A23187 (n = 4). When BRPs were incubated with exogenous arachidonic acid (20 µM) before A23187 stimulation there was also no significant LTC₄/D₄/E₄ production (n = 2). PMNs did not synthesize LTs in the absence of A23187. However, when challenged with the ionophore, there was slight (peak: 5.3 ± 1 pg/µg protein; n = 12), but, significant LT production, possibly the result of a small number (3.3 ± 0.7%; n = 8) of eosinophils contaminating the PMNs.

When PMNs were coincubated with BRPs, basal levels of the LTs were unmeasurable (n = 12). However, stimulation with A23187 elicited a concentration-dependent increase in LTC₄/D₄/E₄ production (fig. 1). At the highest concentration of A23187 tested, BRPs coincubated with PMNs produced approximately seven times more LTs (peak: 35.4 ± 4 pg/µg protein) than PMNs alone.

View larger version (15K): [in this window] [in a new window]   Fig. 1.   Bovine retinal pericytes (BRPs) coincubated with human polymorphonuclear cells (PMNs; 1 × 106 per well) produce significant amounts of sulfidopeptide leukotrienes (LTC4/D4/E4) in response to A23187. Data are mean ± S.E.M. of 9 to 15 observations. *P < .05 when compared to .03 µM A23187.

The lipoxygenase inhibitor, esculetin (1 × 10⁴ M) had no effect on basal or A23187-induced LT production from either BRPs or PMNs (n = 12). In contrast, esculetin (1 × 10⁴ M) reduced (by 63.9 ± 7%) A23187-induced peak production of the sulfidopeptide LTs from BRPs cocultured with PMNs (n = 12, fig. 2).

View larger version (47K): [in this window] [in a new window]   Fig. 2.   Sulfidopeptide leukotrienes (LTC4/D4/E4) production induced by A23187 (1 µM) in BRPs coincubated with human polymorphonuclear cells (PMNs; 1 × 106 per well) in the absence (hatched column) and presence (solid column) of esculetin (1 × 104 M). Data are mean ± S.E.M. of 12 observations. *P < .05, when compared to vehicle control.

In a separate series of experiments, authentic LTA₄ (10 nM) incubated with BRPs elicited the production of LTC₄/D₄/E₄ (18.9 ± 5 pg/µg protein, n = 3).

To determine whether the pericytes exhibited the message for 5-lipoxygenase, total cellular RNA was extracted from the bovine retinal pericytes as well as from human and bovine neutrophils. Northern Blot analysis revealed mRNA for 5-lipoxygenase to be present in human and bovine neutrophils, but no mRNA for 5-lipoxygenase was detected in the RNA obtained from bovine retinal pericytes (fig. 3).

View larger version (27K): [in this window] [in a new window]   Fig. 3.   Northern blot analysis of RNA obtained from human (lane 1) and bovine (lane 3) neutrophils and bovine retinal pericytes (lane 2) revealed message for 5-lipoxygenase to be present in human and bovine neutrophils, but to be absent in the pericytes. Because the 5-lipoxygenase specific probe was constructed from human DNA, the human neutrophils served as positive controls and the bovine neutrophils served to demonstrate that there was cross reactivity with bovine mRNA for 5-lipxygenase.

	Discussion

Top Abstract Introduction Methods Results Discussion References

We demonstrate that pericytes can produce sulfidopeptide LTs, but only if LTA₄ is supplied by a donor cell such as the neutrophil. Pericytes cannot synthesize LTC₄ directly from endogenous or exogenous arachidonic acid, even when stimulated with A23187. Thus, it appears that pericytes, like endothelial cells (Feinmark and Cannon, 1986), vascular smooth muscle cells (Feinmark and Cannon, 1987), type II pneumocytes (Robidoux et al., 1994) and keratinocytes (Iversen et al., 1994), lack the 5-lipoxygenase enzyme, i.e., they are incapable of synthesizing LTA₄. The pericytes do, however, appear to possess the enzymes (LTC₄ synthase, -glutamyl transpeptidase, cysteinylglycine dipeptidase) required for synthesis of the sulfidopeptide LTs. Thus, if provided with LTA₄ either from a cellular source such as the PMN or by the addition of authentic LTA₄, the pericytes is quite capable of producing sulfidopeptide leukotrienes. Northern blot analysis of the RNA obtained from bovine retinal pericytes supported this conclusion in that no mRNA for 5-lipoxygenase was detected in the pericytes, but it was detected in human as well as bovine neutrophils. Passage 1 pericytes were used in this study to avoid loss of enzyme activity that sometimes occurs with multiple passages. Thus, 5-lipoxygenase activity was reported to be lost in a human mast cell line after 10 passages (Macchia et al., 1995). In one additional study, using mouse mast cells, both 5-lipoxygenase activity as well as phospholipase activity were reported to be lost only when the cells attached to plastic (Xu et al., 1993). In our experiments there was no evidence of phospholipase loss in that the pericytes maintained capacity to synthesize prostaglandins (data not shown).

Although it has been postulated that pericytes modulate microvascular tone and permeability, the mechanisms underlying such modulation are unresolved. For many years it has been suggested that, pericytes are contractile cells, which contribute to the regulation of blood flow at the microvascular level (Rouget, 1873; Vimtrup, 1922). Pericytes in vitro contract to a greater extent than do microvascular endothelial cells (Kelley et al., 1987). This suggests that, although endothelial cells contribute to vascular tone, the pericyte is the principal contributor to the regulation of capillary caliber. Moreover, endothelial cells interact with pericytes to affect pericyte contractility in vitro (Dodge et al., 1991). They may do so by donating, for example, endothelin, angiotensin II, prostacyclin or nitric oxide; endothelial-derived mediators which are known to regulate pericyte contractility (Dodge et al., 1991). Endothelial cells (Feinmark and Cannon, 1986) and pericytes (this study) synthesize the sulfidopeptide LTs from LTA₄. It is not yet known whether these LTs alter pericyte contractility. However, it is highly likely that they do as both LTC₄ and LTD₄ induce contraction of glomerular mesangial cells (a cell that is considered to be a type of pericyte; Schlondorff, 1987) in culture (Simonson and Dunn, 1986). Thus, one way in which pericytes may affect microvascular tone and/or permeability is via the production of the sulfidopeptide LTs which may act in either a paracrine or an autocrine manner.

Circulating PMNs coming into contact with an intact microvascular endothelium, may induce the synthesis of sulfidopeptide LTs (Feinmark and Cannon, 1986). These mediators could be released abluminally to alter pericyte contractility. Alternatively, the interactions between pericytes themselves and neutrophils may be potentially very important. In a rat cremaster model of histamine-induced inflammation, neutrophils were shown to invade the interstitial space, where they would be expected to encounter the underlying pericytes (Sims et al., 1990). In addition, if the endothelium is damaged, the neutrophil can make direct contact with the pericyte. In both cases the interaction of pericytes with neutrophils could lead to the production of sulfidopeptide LTs. As mentioned previously, LTC₄ and LTD₄ are powerful inflammatory and contractile agents (Dahlén et al., 1980; Dahlén et al., 1981; Williams and Piper, 1980; Björk et al., 1983; Greenwald et al., 1984). Ultrastructural studies demonstrate that postcapillary venules are highly sensitive to vasoactive-induced increases in permeability (Majno et al., 1961). Moreover, these vessels are suggested to have the highest proportion of pericytes in the vasculature (Tilton, 1991). This again implies that LTs derived from pericytes play an important role in regulating vascular permeability.

`Transcellular biosynthesis' has been postulated to be an important mechanism for regulation of eicosanoid production in several pathophysiological states, notably, acute myocardial infarction, acute lung injury and inflammation (Maclouf et al., 1989). PMN perfusion of rabbit isolated, lungs (Grimminger et al., 1990) or heart (Sala et al., 1993) elicits sulfidopeptide LT production. This is most likely due to the uptake of PMN-derived LTA₄ by endothelial cells. Here, we demonstrate that the interaction between pericytes and PMNs results in the formation of sulfidopeptide LTs and we hypothesize that this may be a potentially important cell-cell interaction in the microcirculation.

In conclusion, these results demonstrate that pericytes are a potentially important source of sulfidopeptide LTs in the microcirculation. Pericytes can convert LTA₄, but not arachidonic acid, to LTC₄. Pericyte incubation with A23187 and PMNs elicited substantial sulfidopeptide LT production, suggesting that LTA₄ is transferred from neutrophils to pericytes. This transcellular LT biosynthesis may play an important role in modulating fluid movement, vascular tone and permeability in the microvasculature, particularly in inflammatory states.