|Home | About | Journals | Submit | Contact Us | Français|
PECAM-1 is a cell adhesion and signaling receptor that is expressed on many hematopoietic cells and at endothelial cell-cell junctions. Accumulating evidence from a number of in vitro and in vivo model systems suggests that PECAM-1 functions to suppress cytokine production and vascular permeability induced by a wide range of inflammatory stimuli. In several of these models of inflammatory disease, endothelial, and not leukocyte or platelet, PECAM-1 has been found to confer protection against inflammatory insult. The mechanism by which endothelial PECAM-1 functions as an anti-inflammatory protein, however, is poorly understood. It has recently been suggested that PECAM-1 exerts its anti-inflammatory effects in endothelial cells by inhibiting the activity of NFκB, a pro-inflammatory transcription factor. To confirm and extend these observations, we examined the effect of engaging, cross-linking, or expressing PECAM-1 on NFκB activation in a variety of human cells. PECAM-1 had no effect on the phosphorylation of the NFκB inhibitory protein, IκBα, on nuclear translocation of NFκB, on suppression of cytokine-induced transcriptional activation of an NFκB luciferase reporter plasmid, or on cytokine-stimulated upregulation of ICAM-1, an NFκB target gene, in endothelial cells. Taken together, these studies strongly suggest that the anti-inflammatory actions of PECAM-1 in endothelial cells are not likely to involve its regulation of NFκB.
PECAM-1 (CD31) is a member of the Ig-superfamily of cell adhesion molecules that is expressed on most cells of the hematopoietic lineage including platelets, monocytes, neutrophils, and certain lymphocyte subsets (1-3). PECAM-1 is also a major constituent of endothelial cell inter-cellular junctions in confluent vascular beds, where it has been shown to play an important role in leukocyte diapedesis (4-6). In addition to its support of this pro-inflammatory event, however, a growing number of studies in C57BL/6 mice have revealed that PECAM-1 can also function prominently in dampening the inflammatory response in a variety of clinically relevant acute and chronic inflammatory conditions, including collagen-induced arthritis (7,8), late-stage autoimmunity (9), autoimmune encephalitis (10), lipopolysaccharide-induced endotoxic shock (11,12), atherogenic diet-induced steatohepatitis (13), and atherosclerosis (14). In the case of LPS-induced endotoxic shock and autoimmune encephalitis, it has been shown through use of bone marrow chimeric animals that endothelial, but not leukocyte, PECAM-1 is largely responsible for protection against excessive inflammation (10,11). The mechanism by which PECAM-1, and specifically endothelial PECAM-1, is protective against exaggerated inflammation, however, is poorly understood.
The NFκB/Rel family of transcription factors is made up of seven proteins – p105, p100, p50, p52, RelA (p65), c-Rel, and RelB – that are important in cell growth, differentiation, inflammation, and survival (15). The NFκB molecule typically consists of a heterodimer (the prototype is RelA and p50) that is sequestered in the cytosol and inhibited by its cytosolic binding partner, IκBα. Upon cellular activation by various stimuli, the phosphorylation and subsequent degradation of IκBα releases the NFκB dimer, allowing it to translocate to the nucleus where it acts as a transcription factor upregulating the expression of numerous pro-inflammatory genes (16). Pro-inflammatory genes turned on by NFκB include acute phase proteins in the liver, endothelial cell adhesion molecules such as ICAM-1 and VCAM-1, and various leukocyte and endothelial cell cytokines that further exacerbate the inflammatory response (15).
A recent study reported that engagement of PECAM-1 results in downregulation of nuclear levels of NFκB in cytokine-stimulated endothelial cells (17). These authors put forward the attractive hypothesis that PECAM-1-PECAM-1 interactions between transmigrating leukocytes and endothelial cells initiate a negative feedback loop that prevents excessive leukocyte recruitment to sites of inflammation by dampening the expression of pro-inflammatory adhesion molecules on the endothelial cell surface (17). Based on these observations, we sought to confirm and extend our understanding of the mechanism by which PECAM-1 functions as an anti-inflammatory signaling receptor in the vascular endothelium. Contrary to this previous report, however, we found using a series of complementary experimental systems that neither PECAM-1 engagement nor crosslinking has an effect on inhibiting NFκB activity, as determined by Western blot for phosphorylated and total IκBα, immunofluorescence for translocation of NFκB subunits, or on binding to NFκB target oligonucleotides by EMSA. We also found that PECAM-1 expression does not inhibit NFκB transcriptional activity in cytokine-stimulated HEK293 cells containing an NFκB luciferase reporter plasmid, or prevent the upregulation of ICAM-1, an NFκB target gene, in cytokine-stimulated endothelial cells. Taken together, these studies demonstrate that the anti-inflammatory actions of PECAM-1 in endothelial cells are likely not due to regulation of NFκB activity.
All primary and secondary antibodies used in cross-linking experiments were tested for endotoxin contamination using the QCL-100 Limulus amebocyte lysate (LAL) kit (Lonza, Walkersville, MD) according to manufacturer’s instructions. Reagents that tested positive for endotoxin contamination were subsequently decontaminated using Detoxi-Gel endotoxin removal columns (Thermo Scientific, Rockford, IL) according to manufacturer’s instructions. Detoxified reagents were below the limit of detection in subsequent assays using the QCL-100 kit.
Passage 3 human umbilical vein endothelial cells (HUVEC) were obtained from human volunteer donors by the Hybridoma and Endothelial Cell Core laboratory at the Blood Research Institute and maintained in RPMI, 10% fetal bovine serum (FBS, Sigma Aldrich, St. Louis, MO), 2 mM L-glutamine, 1 mM sodium pyruvate, 10 mM HEPES, 500 μg/ml gentamicin, 6.45 U/ml heparin (Sigma Aldrich-H3149), and 500 μg/ml endothelial cell growth supplement (BD Biosciences, San Jose, CA). All cell culture reagents were obtained from Mediatech, Inc. (Manassas, VA) unless otherwise specified. Confluent monolayers of HUVEC were stimulated with PBS or IL-1β (1 ng/ml; Peprotech Inc., Rocky Hill, NJ) + platelet activating factor (PAF, 10−10M, Sigma Aldrich) for either 4 or 5 hours (see Results). Cell surface PECAM-1 was then cross-linked with 20 μg/ml Hec7 (a mouse monoclonal antibody (mAb) specific for PECAM-1 Ig Domain 1 that mimics PECAM-1/PECAM-1 homophilic interactions - kind gift of Dr. William Muller, Northwestern University), followed by addition 30 minutes later of 5 μg/ml of goat anti-mouse IgG (Jackson Immunoresearch, West Grove, PA). Cells were then lifted with trypsin/EDTA, washed, and incubated on ice for 15 minutes in Buffer A (200 μl of 10 mM KCl, 10 mM Hepes, pH 7.9, 0.1 mM EDTA, pH 8.0, 0.1 mM EGTA, 1 mM dithiothreitol) with 1:100 Protease Inhibitor Cocktail Set I (CalBiochem, San Diego, CA) before addition of NP-40 to a final concentration of 0.5%. After vortexing for 10 seconds, the mixture was centrifuged at 20,000 g for 30 seconds, the resulting nuclear pellet resuspended in 40 μl of Buffer B (20 mM HEPES, pH 7.9, 0.4 M NaCl, 1 mM EDTA, pH 8.0, 1 mM EGTA, 1 mM dithiothreitol, 1:100 Protease Inhibitor Cocktail Set I) and incubated on ice for an additional 20 minutes. Nuclei were finally pelleted by centrifugation at >20,000 g for 10 minutes, and the protein concentration of the resulting supernatant determined using the BCA assay (Thermo Scientific). Glycerol was then added to a final concentration of 10% and aliquots stored at −80 °C until use.
Forward (5′-AGTTGAGGGGACTTTCCCAGGC-3′) and reverse (5′-GCCTGGGAAAGTCCCCTCAACT-3′) oligonucleotides containing the consensus NFκB binding sequence were obtained from Integrated DNA Technologies (San Diego, CA). Ten μg of each were annealed for 5 minutes at 90°C in 1× T4 kinase buffer (GE Healthcare Life Science, Piscataway, NJ) in a final volume of 100 μl, cooled to room temperature, and finally incubated at 37°C for 1 hour to create the double-stranded probe. Double-stranded Oct1 oligonucleotide was obtained from Santa Cruz Biotechnologies and used as a loading control. Both probes were labeled with (γ-32P) ATP using a DNA 5′ End Labeling Kit (Promega, Madison, WI) according to manufacturer’s instructions, and had specific activities of > 2 × 109 cpm/μg. Nuclear extracts (15 μg) were incubated with 2 μg of poly(dI-dC) and 10 mM Tris-HCl buffer, pH 7.5, containing 50 mM NaCl, 0.5 mM EDTA, 1 mM dithiothreitol, 10% (w/v) glycerol, and labeled probe (> 2 × 106 cpm/reaction). The reactions were electrophoresed on a 5% polyacrylamide gel containing 0.5× TBE and 2.5% glycerol, dried onto Whatman paper, and then exposed by autoradiography. For band intensity determination, the dried gel and film were aligned, bands were traced and cut out of the dried gel, and put in scintillation fluid. β-particle emission counts of individual bands were measured on a Wallac 1410 liquid scintillation counter.
Passage 3 HUVEC were grown to confluence and stimulated with 1 ng/ml IL-1β+PAF (10−10M) in the presence of 20 μg/ml of Hec7 or PECAM 1.3 mAbs. After 10, 30, or 60 minutes, cells were lysed in ice-cold Lysis Buffer (20 mM Tris, 150 mM NaCl, 1% Triton X-100, 1 mM EDTA, 10 mM NaF), containing 1:100 Phosphatase Inhibitor Cocktail Set I (CalBiochem). Lysates were incubated at 4°C for 30 minutes and spun down at >20000g for 20 minutes at 4°C. Supernatants were removed and 30 μg of each lysate was loaded onto a 10% SDS/PAGE gel, electrophoresed, and transferred to an Immobilon PVDF membrane. Membranes were blocked with 5% milk and incubated with anti-phospho-IκBα (1:1000, Cell Signaling Technology #9246, Danvers, MA) and HRP-conjugated anti-mouse IgG (Jackson Immunoresearch). After image detection, membranes were washed and incubated with anti-IκBα (1:1000, Cell Signaling #9242) and HRP-anti-rabbit IgG, imaged, stripped, and re-probed with anti-α-tubulin (1:1000, Cell Signaling #2125) and HRP-anti-rabbit IgG. Images were obtained on the Kodak Image Station 2000R or an x-ray film developer. Band densitometry quantitation was measured using Kodak Molecular Imaging software.
Passage 3 HUVEC were plated and grown to confluence on gelatin-coated 8-chamber glass slides (BD Biosciences). Confluent monolayers of HUVEC were stimulated with PBS or IL-1β (1 ng/ml) + PAF (10−10M) for 1, 3, or 5 hours. One hour prior to the end of stimulation for each timepoint, cell surface PECAM-1 was cross-linked with 20 μg/ml of mAb Hec7 or mAb PECAM-1.3 or PECAM-1/IgG chimeric fusion protein, followed 30 minutes later by the addition of 5 μg/ml of goat anti-mouse IgG (for Hec7 and PECAM 1.3) or 5 ug/ml goat anti-human IgG (for PECAM-1/IgG). Cells were fixed with 2% paraformaldehyde and permeabilized with 0.2% Triton X-100. Cells were blocked with 3% BSA and then incubated sequentially with anti-NFκB p65 (1:200, Santa Cruz #sc109), Texas Red goat anti-rabbit (1:100), and FITC-β-catenin (1:65, Cell Signaling #2849). Nuclei were counterstained with DAPI and slides were mounted with ProLong Gold antifade reagent (Invitrogen, Carlsbad, CA). Images were obtained using a 40× oil objective and lasers at 405, 488, and 559 nm on an Olympus FluoView FV1000 MPE microscope (Center Valley, PA, USA). Z-stack fluorescent images were acquired at 4 μs per pixel at optimal stepwise progression through the whole thickness of the monolayer. Images were reconstructed and analyzed with SlideBook 5.0 imaging software. All images were uniformly background-subtracted, and a universal mask was generated to identify nuclei based on the DAPI channel. The mask excluded all objects <500 pixels and along the edge of the image. Nuclear levels of NFκB were obtained by measuring the mean fluorescence intensity of the Texas Red (NFκB p65) channel within the mask throughout the whole z-stack image. Three separate images from each experimental group were analyzed, and graphs depict the mean fluorescence intensity of nuclear NFκB for all nuclei analyzed. For representative images, a projection image of combined z-stacks was created using SlideBook 5.0 software. Images were uniformly background subtracted, and brightness and contrast was adjusted in PowerPoint separately, but uniformly, for images at each timepoint.
HEK293 cells were transfected with pcDNA3.1 vector and pNifty2-Luc NFκB luciferase reporter plasmid (Invitrogen, San Diego, CA) using Superfect transfection reagent (Qiagen, Valencia, CA). Double stable clones were selected with G418 and Zeocin. Expressing clones were plated onto 96 well tissue culture plates and grown to 80% confluence. Cells were co-transfected with 0.2 μg pWPT-GFP (Addgene #12255) and either 0.2 μg of plasmid containing dsRed or human PECAM-1 cDNA at a 4× lipid to DNA ratio with Lipofectamine LTX transfection reagent (Invitrogen) according to manufacturer’s instructions. Three days later, cells were stimulated with IL-1β (1 ng/ml) + PAF (10−10M) or TNFα (1 ng/ml; R & D Systems, Minneapolis, MN) + PAF for 6 hrs. eGFP fluorescence was measured using a Victor2 multi-label photon counter. Luciferin salt (Xenogen; kind gift of Dr. Mike Dwinnell, Medical College of Wisconsin) was subsequently added to wells, and luciferase activity was measured on a Victor2 multi-label photon counter. After luciferase readings, cells were lifted and eGFP fluorescence was analyzed on an LSRII flow cytometer to ensure similar transfection efficiency and eGFP fluorescence intensity (data not shown). Since transfection efficiency and eGFP fluorescence intensity was similar for all wells, we divided the relative light units (RLU) from luciferase readings by the eGFP readings obtained on the Victor2 multi-label photon counter to normalize for cell number. For western blots assessing PECAM-1 expression, lysates were made from transfected cells, 40 μg lysate was loaded onto an SDS-PAGE gel, electrophoresed, transferred to membranes, and blots were imaged on the Kodak Image Station 2000R as described above. Sew 32-34 (1 μg/ml, a rabbit polyclonal antibody directed against human PECAM-1) and anti-α-tubulin (1:1000) were used for blotting.
Passage 3 HUVEC were grown to confluence and stimulated with IL-1β (1 ng/ml) + PAF (10−10M). Four hours after addition of cytokines, 20 μg/ml of mAb Hec7 or mAb PECAM-1.3 or PECAM-1/IgG chimeric fusion protein was added, allowed to incubate for 30 minutes, and then cross-linked by addition of 5 μg/ml species-specific secondary antibody. Two hours later (6 hours total), cells were lifted with trypsin/EDTA and double-stained for ICAM-1 and PECAM-1 using FITC-ICAM-1 (1:100, Santa Cruz #sc-107) and Alexa647-PECAM-1.3 (50 μg/ml). In other experiments, HUVEC were stimulated with 10 ng/ml TNFα, and 10 minutes after addition of TNFα, 50 μg/ml of mAb PECAM-1.3 was added, allowed to incubate for 10 minutes, and then cross-linked by addition of 100 μg/ml goat anti-mouse IgG. Six hours later, cells were lifted with trypsin/EDTA double-stained for ICAM-1 and PECAM-1 using FITC-ICAM-1 (1:100, Santa Cruz #sc-107) and Alexa680-PECAM-1.3 (50 μg/ml). In some experiments, PECAM-1 expression was knocked down using lentivirus constructs encoding PECAM-1 siRNA (18) before stimulation with 1 ng/ml IL-1β + PAF (10−10M) or TNFα (20 ng/ml) for 6 or 24 hours, after which cells were lifted with trypsin/EDTA and double-stained as above. FACS was performed using an LSRII flow cytometer. Alexa647-PECAM-1.3 and Alexa680-PECAM 1.3 were created by labeling mAb PECAM-1.3 with the Alexa Fluor 647 and Alexa 680 monoclonal antibody labeling kits (Molecular Probes, Carlsbad, CA) according to manufacturer’s instructions.
Results, where applicable, are expressed as mean ± standard deviation. Statistical analysis was performed on GraphPad Prism 5 software.
To confirm an earlier report (17) that engagement of PECAM-1 functions to suppress NFκB activity in the vascular endothelium, we stimulated HUVEC with IL-1β+PAF for four hours, and then crosslinked PECAM-1 with mAb Hec7+anti-mouse IgG for an additional hour, either in the presence or absence of continuing exposure to the activating cytokines. As shown in Figure 1, antibody-mediated engagement of PECAM-1 did not decrease cytokine-initiated translocation of NFκB to the nucleus, as measured by gel-shift analysis, either under the original conditions described by Cepinskas et al. (17) in which IL-1β+PAF was withdrawn prior to PECAM-1 engagement (compare lane 4 with lane 5), or under the more physiological conditions recently described by this group (19) in which these activating cytokines were allowed to remain present during the PECAM-1 cross-linking process (compare lane 2 with lane 3).
Because these findings differed from those previously reported (17,19), the ability of PECAM-1 engagement to suppress cytokine-induced nuclear translocation of NFκB was also examined by immunofluorescence microscopy. As shown in Figure 2, engagement of cell surface PECAM-1 with mAb Hec7 or a bivalent PECAM-1/IgG chimeric fusion protein, both of which bind to extracellular Ig domain 1 (20), had no effect on the translocation of NFκB p65 to the nucleus after 1, 3, or 5 hours of IL-1β+PAF stimulation. A second anti-PECAM-1 mAb, PECAM-1.3, had only a minor effect, and that only at earlier time points. These data, like those obtained in Figure 1, suggest that engagement of endothelial cell surface PECAM-1 following cytokine stimulation has little, if any, biologically-significant ability to affect translocation of NFκB to the nucleus.
To examine the possibility that engagement of PECAM-1 might be acting upstream of NFκB to suppress its activation, we examined the effect of monoclonal antibody-mediated engagement of PECAM-1 on IL-1β+PAF-induced phosphorylation and degradation of IκBα – a cytosolic suppressor of NFκB activation. Binding of mAbs Hec7 and PECAM 1.3 did not inhibit either the initial phosphorylation (Fig. 3a and 3b), degradation (Fig. 3a and 3c), or rate of antigen recovery (Fig. 3a and 3c, Supplemental Fig. 1) of IκBα in HUVEC. Taken together with the findings of Figures 1 and and22 above, we conclude that engagement of PECAM-1 has negligible effects on cytokine-stimulated activation of NFκB.
Though PECAM-1 engagement per se has negligible effects on NFκB activity (Figures 1--33 above), it is possible that expression of PECAM-1 alone might be sufficient to modulate NFκB responsiveness. To determine whether PECAM-1 expression affects the transcriptional activity of NFκB, we transfected either a dsRed control plasmid or a cDNA encoding human PECAM-1 into HEK293 cells that had been stably transfected with an NFκB-responsive luciferase reporter plasmid (Fig. 4a). As shown in Figure 4b, both dsRed-transfected and PECAM-1-transfected cells responded to exposure to IL-1β+PAF and TNFα+PAF with an increase in NFκB transcriptional activity. Expression of PECAM-1, however, had no effect on the extent of NFκB activation following stimulation with either inflammatory cytokine combination. As such, it would appear that expression of PECAM-1 is not able to inhibit transcriptional activation of NFκB.
ICAM-1 serves as an important counter-receptor for leukocyte integrins during the leukocyte adhesion cascade (21), and becomes markedly upregulated in an NFκB-dependent manner on the surface of cytokine-activated endothelial cells (15). To determine whether PECAM-1 engagement might be sufficient to dampen the upregulation of ICAM-1 in cytokine-activated endothelial cells, we performed flow cytometric analysis of HUVEC in which PECAM-1 was cross-linked with anti-PECAM-1 mAbs or a bivalent PECAM-1/IgG chimeric protein. As shown in Figure 5a, engagement of PECAM-1 by either mAbs Hec7 or PECAM 1.3, or with bivalent PECAM-1/IgG, was not able to suppress IL-1β+PAF-induced upregulation of ICAM-1 after 6 hours of cytokine stimulation. Similar results were seen after 24 hours of IL-1β+PAF stimulation (data not shown). Likewise, engagement using mAb PECAM 1.3 also showed no effect on TNFα-induced upregulation of ICAM-1 (Fig. 5b). siRNA-mediated silencing of PECAM-1 also failed to have an effect on cytokine-induced ICAM-1 expression in these cells (Fig. 5c & 5d). Taken together, these studies demonstrate that neither engagement nor expression of PECAM-1 is able to modulate NFκB-dependent responses of the inflamed endothelium.
PECAM-1 is a well-studied cellular adhesion and signaling receptor that plays an important role in supporting leukocyte diapedesis during the leukocyte adhesion cascade (4,21). In contrast to this pro-inflammatory effect, PECAM-1 has been shown in a number of situations to function as an ITIM-containing inhibitory receptor capable of dampening cellular activation events in lymphocytes (9,22,23), mast cells (24), and platelets (25-28). Numerous reports have also demonstrated an anti-inflammatory role for PECAM-1 in well-established acute and chronic inflammatory disease models. For example, mice expressing PECAM-1 produce lower levels of inflammatory cytokines (7,11-13), exhibit enhanced vascular barrier protection (10-12), and paradoxically, accumulate fewer leukocytes at sites of inflammation (10-13). The mechanisms by which PECAM-1 serves to confer protection in inflammation, however, and how it regulates these aspects of the inflammatory response are still poorly understood.
One intriguing possibility might be that PECAM-1 inhibits signaling of a canonical pro-inflammatory signaling pathway, thereby dampening cellular activation and inflammation. Cepinskas, et al. reported that PECAM-1 engagement on the endothelial cell surface was able to decrease nuclear levels of NFκB, which they hypothesized might be responsible for initiating a negative feedback loop to dampen further cellular activation, namely prolonged expression of endothelial adhesion molecules (17). Since endothelial PECAM-1 likely confers much of the protection against excessive inflammation, at least in models of LPS-induced endotoxic shock (11,12) and murine autoimmune encephalitis (10), we hypothesized, like Cepinskas, et al., that endothelial PECAM-1 might function to dampen inflammatory activation of endothelial cells through inhibition of NFκB.
To test this hypothesis, we engaged and cross-linked PECAM-1 on the surface of IL-1β+PAF-stimulated HUVEC using two different anti-PECAM-1 mAbs, as well as a homophilically-binding PECAM-1/IgG chimeric protein construct. In preliminary experiments, we found no detectable decrease in nuclear translocation of NFκB, as detected by EMSA, when we engaged PECAM-1 with mAbs PECAM-1.3 and PECAM-1.2, or with PECAM-1/IgG (data not shown). It was subsequently revealed (19) that the observation that PECAM-1 modulated NFκB activity had been made under conditions in which the stimulatory cytokines IL-1β+PAF were removed prior to antibody addition. We therefore performed a side-by-side comparison of the effects of PECAM-1 engagement either in the continuous presence (our protocol) or following removal (17,19), of IL-1β+PAF. Although the amount of NFκB detected in the nucleus decreased following cytokine removal (Figure 1), engaging PECAM-1 had no discernable additional effect, even when using the same anti-PECAM-1 mAb (Hec7) that had been used in the previously published experiments (17). Similarly, addition of anti-PECAM-1 mAbs or PECAM-1/IgG had no effect on the rate or extent of phosphorylation and degradation of IκBα (Figure 3) or on upregulation of the NFκB-inducible leukocyte adhesion molecule, ICAM-1, in IL-1β+PAF- or TNFα-stimulated HUVEC (Figure 5a & 5b). Cross-linking of cell surface PECAM-1 with mAb PECAM 1.3 did cause a slight decrease in the amount of translocated NFκB p65 at 1 and 3 hours of IL-1β+PAF stimulation (Figure 2), which would support findings by Cepinskas, et al.(17). The small increment of this decrease, however, coupled with the fact that two other PECAM-1-specific cross-linking reagents had no effect on the translocation of NFκB (Figure 2), and that PECAM 1.3-mediated cross-linking of PECAM-1 had no effect on cytokine-induced upregulation of ICAM-1 (Figure 5a & 5b), strongly suggest that engaging and cross-linking PECAM-1 homophilically is unable to dampen cytokine-mediated activation of the NFκB pathway in the vascular endothelium. The possibility remains, however, that during the process of neutrophil transmigration, PECAM-1 becomes engaged and activated in a manner that has not been adequately mimicked by the reagents used in this study, and that such interactions might initiate anti-inflammatory signaling pathways mediated by PECAM-1. Given the recent identification of CD177 as a high-affinity heterophilic ligand for PECAM-1 (29), the possibility that engagement of PECAM-1 in this manner might be able to send inhibitory signals into the cell would be an interesting line of future investigation.
Despite the inability of PECAM-1 homophilic engagement to suppress NFκB activation, we employed two complementary approaches to examine the possibility that PECAM-1 expression per se might be sufficient to dampen NFκB activation in response to inflammatory mediators. In the first, we compared NFκB-mediated luciferase activity in cytokine stimulated PECAM-1-positive versus PECAM-1-negative HEK293 cells that harbored an NFκB-responsive luciferase reporter gene (Figure 4). In the second, we examined the effect of siRNA-induced PECAM-1 silencing on the upregulation of ICAM-1 expression in cytokine-stimulated HUVEC (Figure 5c & 5d). In neither case was PECAM-1 expression found to have an inhibitory effect in the presence of either IL-1β+PAF or TNFα. As such, we conclude that PECAM-1 expression, like PECAM-1 engagement, is not a regulator of cytokine-induced activation of NFκB in endothelial cells.
In summary, PECAM-1 appears to play a number of prominent, though sometimes opposing, roles in regulating specific components of the inflammatory response. On the positive, pro-inflammatory side, PECAM-1 supports leukocyte transendothelial migration, and functions as a positive regulator of NFκB in atheroprone areas of the vasculature where endothelial cells are subjected to oscillatory, disturbed fluid shear (30,31). On the anti-inflammatory side, PECAM-1 plays a role in maintaining the barrier function of endothelial cell-cell junctions (10-12) and suppresses cytokine production (7,11-13). In this regard, Rui, et al. (32) reported that heterophilic engagement of macrophage PECAM-1 with a CD38-Fc fusion protein depressed LPS-induced pro-inflammatory cytokine and type I IFN production by inhibiting JNK, IRF3, and NFκB. It would seem, therefore, that the ability of PECAM-1 to influence inflammation in general, and NFκB in particular, may depend on how PECAM-1 is engaged, the cell type that it resides in, and the environmental conditions to which that cell is subjected. Evidence that cells employ PECAM-1 in both a situation- and site-specific manner to “fine-tune” their response to cellular stress can be found in three recent reports, each of which examined the role of PECAM-1 in atherosclerosis, a complex disorder whose development and progression is influenced by both rheological factors, the presence of inflammatory stimuli, and the ease at which leukocytes migrate into the vessel wall. Thus, the pro-inflammatory properties of endothelial cell PECAM-1, likely acting as a sensor of shear-stress that activates of NFκB, was found to dominate and promote atherosclerosis in the aortic arch (14,31,33), while the ability of PECAM-1 to suppression inflammation, namely the production of pro-inflammatory cytokines (7,11-13) and reactive oxygen species (13,34), appear to play a more prominent role within the descending aorta (14). The molecular details of how PECAM-1 supports anti-inflammatory signaling in platelets, leukocytes, and endothelial cells to attenuate inflammation are not known. We are currently in the process of analyzing real-time PCR-based microarrays to generate PECAM-1-regulated candidate genes that might confer such protection.
We are grateful to Dr. William Muller (Northwestern University) for his helpful comments, and for supplying the Hec7 mAb, and to Dr. Hartmut Weiler (BloodCenter of Wisconsin) for his critical review of the manuscript. We would also like to acknowledge Dr. Yuhong Chen and Dr. Demin Wang (BloodCenter of Wisconsin), Dr. Mike Dwinnell (Medical College of Wisconsin), and Tracey Berg and Dr. Carol Williams (Medical College of Wisconsin) for technical expertise with EMSA, luciferase reporter, and immunofluorescence experiments, respectively.
1This work was supported by Predoctoral Fellowship Award 0810167Z (to JRP) from the Midwest Affiliate of the American Heart Association, and by grant HL-40926 (to PJN) from the National Heart, Lung, and Blood Institute of the National Institutes of Health.
Category: Hemostasis, thrombosis, and vascular biology