Reduced leukocyteendothelial cell interactions in the inflamed microcirculation of macrophage migration inhibitory factordeficient mice.код для вставкиСкачать
ARTHRITIS & RHEUMATISM Vol. 50, No. 9, September 2004, pp 3023–3034 DOI 10.1002/art.20470 © 2004, American College of Rheumatology Reduced Leukocyte–Endothelial Cell Interactions in the Inflamed Microcirculation of Macrophage Migration Inhibitory Factor–Deficient Mice Julia L. Gregory,1 Michelle T. Leech,1 John R. David,2 Yuan H. Yang,1 April Dacumos,1 and Michael J. Hickey1 space were also reduced significantly in MIFⴚ/ⴚ mice. In each of these models, the level of P-selectin– dependent rolling was reduced in MIFⴚ/ⴚ mice. Despite this, no difference in P-selectin expression was observed following LPS treatment. However, microvascular shear forces were elevated in MIFⴚ/ⴚ mice, raising a possible mechanism to explain the reduced interactions in these animals. Conclusion. MIFⴚ/ⴚ mice consistently displayed a reduction in P-selectin–dependent rolling, suggesting that MIF exerts proinflammatory effects, in part, via the promotion of P-selectin–mediated rolling. Together, these data indicate that MIF promotes interactions between leukocytes and endothelial cells, thereby enhancing the entry of leukocytes into sites of inflammation. Objective. Macrophage migration inhibitory factor (MIF) is a proinflammatory cytokine with established roles in a range of inflammatory conditions. However, it is not known whether MIF influences inflammation via the direct promotion of leukocyte– endothelial cell interactions. Therefore, the aim of these experiments was to investigate the ability of MIF to regulate leukocyte–endothelial cell interactions in the inflamed microvasculature. Methods. Intravital microscopy was used to examine postcapillary venules in the cremaster muscle and synovium of wild-type and MIF ⴚ/ⴚ mice. Leukocyte–endothelial cell interactions (rolling, adhesion, emigration) were compared under a range of inflammatory conditions. Results. In cremasteric postcapillary venules of MIFⴚ/ⴚ mice, lipopolysaccharide (LPS)–induced leukocyte rolling, adhesion, and emigration were significantly reduced relative to that in wild-type mice. Similar responses were observed in response to tumor necrosis factor ␣ and histamine. Examination of the synovial microvasculature following exposure to carrageenan revealed that leukocyte rolling and adhesion in synovial postcapillary venules and leukocyte entry into the joint Macrophage migration inhibitory factor (MIF) was originally identified as an “activity” present in the supernatant of cultured T cells following activation, which could prevent the random migration of macrophages in vitro. This observation led to the hypothesis that MIF could promote the retention of these leukocytes in sites of inflammation (1,2). The functions of this proinflammatory cytokine are now recognized to extend well beyond those revealed in these original experiments. At a cellular level, MIF has well-characterized activating effects on macrophages, inducing responses such as cytokine release, increased phagocytic activity, and augmentation of nitric oxide production (3–6). Furthermore, there is now a growing body of evidence that MIF plays a central role in promoting inflammatory responses. MIF expression is up-regulated in human rheumatoid arthritis (RA) as well as a number of animal models of inflammation, including endotoxemia/sepsis, Supported by a project grant from the National Health and Medical Research Council of Australia. Dr. Hickey is a National Health and Medical Research Council R. D. Wright Fellow. 1 Julia L. Gregory, BSc, Michelle T. Leech, MBBS, PhD, Yuan H. Yang, PhD, April Dacumos, BSc, Michael J. Hickey, PhD: Centre for Inflammatory Diseases, Monash University, Victoria, Australia; 2 John R. David, MD, PhD: Harvard School of Public Health, Boston, Massachusetts. Address correspondence and reprint requests to Michael J. Hickey, PhD, Centre for Inflammatory Diseases, Monash University Department of Medicine, Monash Medical Centre, 246 Clayton Road, Clayton, Victoria 3168, Australia. E-mail: michael.hickey@med. monash.edu.au. Submitted for publication December 7, 2003; accepted in revised form May 12, 2004. 3023 3024 delayed-type hypersensitivity (DTH), and adjuvantinduced arthritis (7–9). In addition, studies using either MIF blockade or MIF-deficient (MIF⫺/⫺) mice have repeatedly shown attenuation of inflammatory responses in animal models of arthritis, DTH, endotoxemia, colitis, and experimental allergic encephalomyelitis (EAE) (10–17). A number of potential mechanisms for these apparent proinflammatory actions have been identified, including promotion of monocyte Toll-like receptor 4 (TLR-4) expression, increased macrophage survival via inhibition of p53, and counterregulation of the antiinflammatory effects of glucocorticoids (8,18– 20). However, the capacity of MIF to directly regulate the key inflammatory mechanism common to each of these models, i.e., leukocyte recruitment, has not been directly assessed. For leukocytes to be recruited to sites of inflammation, they must undergo a sequence of interactions with the endothelium of postcapillary venules. Initially, leukocytes tether and roll along the endothelial surface until chemoattractant stimuli cause the rolling leukocytes to adhere to the endothelial surface. Some of these adherent leukocytes detach and return to the circulating pool. However, leukocytes that remain adherent are then able to emigrate into the surrounding tissue (21). These steps are mediated by adhesion molecules found on both leukocytes and endothelial cells (22,23). The potential for MIF to directly promote these interactions has not been examined, although there is some evidence to support this possibility. In a model of pulmonary Pseudomonas aeruginosa infection, MIF⫺/⫺ mice displayed reduced neutrophil accumulation (14). In addition, endotoxemia-induced neutrophil accumulation in the lung is reduced after treatment with anti-MIF antibody (24). Although these studies suggest that one of the actions of MIF in inflammatory responses is to promote leukocyte recruitment to affected tissues, this possibility has never been directly assessed. Therefore, the aim of this study was to determine whether MIF promotes leukocyte–endothelial cell interactions during inflammatory responses. To examine this issue, we used intravital microscopy to observe inflamed postcapillary venules of the cremaster muscle and synovium in both wild-type and MIF⫺/⫺ mice, allowing the direct assessment of leukocyte–endothelial cell interactions. In these experiments, MIF⫺/⫺ mice displayed significant reductions in leukocyte rolling and adhesion, as well as reduced entry of leukocytes into the site of inflammation. These observations indicate that one of the mechanisms whereby MIF may contribute to inflammatory responses is through the promotion of GREGORY ET AL leukocyte–endothelial cell interactions in the microcirculation. MATERIALS AND METHODS Animals. MIF⫺/⫺ mice were generated via homologous recombination in J1 embryonic stem cells as described previously (14). They were maintained on a mixed background of 129/Sv ⫻ C57BL/6. Wild-type mice of the same background were bred from MIF⫹/⫹ littermates and used as controls. Antibodies and reagents. Antibodies used for this study were RB40.34, a monoclonal antibody (mAb) against murine P-selectin (20 g/mouse; BD Biosciences, San Diego, CA); RME-1, a mAb against rat and mouse E-selectin (100 g/mouse) and RMP-1, a mAb against rat and mouse P-selectin (100 g/mouse) (both generously provided by Dr. Andrew Issekutz, Dalhousie University, Halifax, Nova Scotia, Canada); and 6C7.1, a mAb against murine vascular cell adhesion molecule 1 (VCAM-1) (90 g/mouse; hybridoma generously provided by Drs. Dietmar Vestweber and Britta Engelhardt, Max Planck Institut, Muenster, Germany). The doses of all function-blocking antibodies used have been shown to be effective in blocking their specific target molecules in vivo (25,26). A110-1 (rat anti–keyhole limpet hemocyanin [anti-KLH]; BD Biosciences) was used as a control mAb in adhesion molecule expression experiments. RB6-8C5 (anti– Gr-1), FA-11 (anti-CD68), and KT3 (anti-CD3) were purified from hybridoma supernatants. Alexa Fluor protein labeling kits were purchased from Molecular Probes (Eugene, OR). Tumor necrosis factor ␣ (TNF␣) was purchased from R&D Systems (through Bioscientific, Gymea, New South Wales, Australia). Local administration of lipopolysaccharide (LPS), TNF␣, and carrageenan. To observe the effects of MIF on the cremasteric microcirculation, LPS in 200 l of saline was given via intrascrotal injection, adjacent to the right cremaster muscle. The cremaster muscle was prepared for intravital microscopy 4 hours after injection, allowing time for significant leukocyte recruitment to occur. Pilot experiments using doses between 1 ng and 100 ng of LPS revealed that 10 ng induced a consistent increase in leukocyte rolling, adhesion, and emigration without compromising microvascular blood flow. A similar inflammatory response model was established in wild-type mice using TNF␣. The procedure for injection and the time course of these experiments were the same as for the LPS model. Pilot experiments using doses between 50 ng and 500 ng of TNF␣ revealed that 50 ng of TNF␣ induced consistent increases in leukocyte recruitment that were similar to those observed previously for the LPS model. Carrageenan administration was used as a model of joint inflammation, as previously described (27). Briefly, 20 l of carrageenan (1% in 0.5% carboxymethylcellulose in saline) was injected into the joint using a 30-gauge needle. Sixty minutes after injection, the synovial microvasculature was prepared for intravital microscopy as described below. Intravital microscopy. Cremaster muscle preparation. Intravital microscopy of the mouse cremaster muscle was performed as previously described (28). Briefly, mice were anesthetized using a combination of 10 mg/kg of xylazine (Troy LEUKOCYTE INTERACTIONS IN THE MICROCIRCULATION OF MIF⫺/⫺ MICE Laboratories, Smithfield, New South Wales, Australia) and 200 mg/kg of ketamine hydrochloride (Pfizer, West Ryde, New South Wales, Australia) administered via intraperitoneal injection. The left jugular vein was cannulated for administration of additional anesthetic and antibodies. The mouse was then placed on a thermo-controlled heating pad to regulate the core temperature to 37°C. The cremaster muscle was dissected free of surrounding tissues and exteriorized onto an optically clear viewing pedestal. The muscle was cut longitudinally with a cautery and held flat against the pedestal by attaching silk sutures to the edges of the tissue. The muscle was then superfused with bicarbonate buffered saline (pH 7.4; at 37°C) and covered with a coverslip that was held in place with vacuum grease. An intravital microscope (Axioplan 2 Imaging; Carl Zeiss Australia, Carnegie, Victoria, Australia) with a 20⫻ objective lens (20⫻/0.40 numerical aperture) and 10⫻ eyepiece was used to observe the cremasteric microcirculation. A color video camera (Sony SSC-DC50AP; Carl Zeiss Australia) was used to project the images onto a calibrated monitor (Sony PVM-20N5E), and the images were recorded for playback analysis using a videocassette recorder (Panasonic NV-HS950; Klapp Electronics, Prahran, Victoria, Australia). We selected 1–3 venules (25–40 m in diameter) for recording in each experiment. Leukocyte rolling flux, rolling velocity, adhesion, and emigration as well as the venular diameter (Dv) were quantitated off-line using playback analysis, as previously described (28,29). The centerline red blood cell (RBC) velocity (VRBC) in the vessel was measured with an optical Doppler velocimeter (Microcirculation Research Institute, Texas A&M University, College Station, TX), and the mean RBC velocity (Vmean) was determined as the VRBC/1.6. The venular wall shear rate (␥) was calculated based on the Newtonian definition: ␥ ⫽ 8(Vmean/Dv) (28). Histamine-induced rolling. To assess the role of MIF in regulating histamine-induced rolling, the cremaster muscle of an untreated mouse was exteriorized. Thirty minutes was allowed for the initial elevated level of rolling to abate. At this point, when rolling was equivalent in wild-type and MIF⫺/⫺ mice, histamine (100 M in superfusion buffer) was superfused over to the cremaster muscle continuously for 60 minutes, and changes in rolling during the 60-minute period were assessed. This concentration of histamine has been shown previously to cause a rapid, P-selectin–dependent increase in leukocyte rolling in rat mesenteric postcapillary venules (30). Knee joint preparation. The murine synovial microcirculation of the knee joint was examined as previously described (31). Briefly, the mouse was anesthetized, placed on a heating pad as described above, and a catheter was inserted into the tail vein to allow administration of additional anesthetic and fluorescent markers. The left hind limb was immobilized with the knee joint in slight flexion. The skin overlying the knee joint was incised and resected, and the patellar tendon was mobilized and divided distally. The patellar tendon and knee joint were carefully removed, exposing the intraarticular synovial tissue. The exposed tissue was flushed with 5 ml of sterile saline, then immersed in a saline-filled chamber, and covered with a coverslip. The synovial microvasculature was examined microscopically using a 20⫻ immersion lens (20⫻/0.50 WI). To visualize leukocytes, mice were injected intravenously with 50 3025 l of 0.05% rhodamine 6G (Sigma-Aldrich, St. Louis, MO) immediately prior to microscopy, as described previously (32). Rhodamine 6G–associated fluorescence was visualized by epiillumination at 510–560 nm, using a 590-nm emission filter. Microscopic images of the synovial microcirculation were projected onto a monitor using a low-light video camera (Dage-MTI IR-1000; Sci-Tech, Preston South, Victoria, Australia), and rolling and adhesion parameters were analyzed as for the cremaster muscle. Quantitation of MIF messenger RNA (mRNA) via real-time polymerase chain reaction (PCR). Real-time PCR was used to measure MIF mRNA in the cremaster muscle. RNA was extracted from tissue samples via homogenization in TRIzol (Gibco BRL, Grand Island, NY). Total RNA (1 g) was reverse transcribed using Superscript II reverse transcriptase (Gibco BRL) and oligo(dT)12-18. PCR amplification was performed on a LightCycler (Roche Diagnostics, Castle Hill, New South Wales, Australia) using SYBR Green I, as previously described (33). Murine MIF and ␤-actin PCR products were used as assay standards. Amplification (40 cycles) was carried out in a total volume of 10 l containing 1 l of dNTP mixture, Taq polymerase, SYBR Green I dye, and the following primers at 3 pM: MIF 5⬘-TGACTTTTAGCGGCACGAAC-3⬘ and 5⬘-GACTCAAGCGAAGGTGGAAC-3⬘, and ␤-actin 5⬘-TGTCCCTGTATGCCTCTGGT-3⬘ and 5⬘-GATGTCACGCACGATTTCC-3⬘. Melting curve analysis and agarose gel electrophoresis were performed at the end of each PCR reaction. The level of MIF mRNA expression was quantitated and expressed relative to ␤-actin expression. The results are presented as the ratio of mRNA expression in LPS-treated mice to that in untreated animals. Quantitation of P-selectin expression in the cremaster muscle. P-selectin expression in the cremaster muscles of LPS-treated wild-type and MIF⫺/⫺ mice was quantitated as previously described (29). Briefly, Alexa 488–conjugated RMP-1 (detected via epiillumination at 450–490 nm, with a 515-nm emission filter) was used to label P-selectin, and Alexa 594–conjugated anti-KLH mAb (detected using a 530–585-nm excitation filter, with a 615-nm emission filter) was used as a nonspecific control IgG. Images were visualized using a SIT video camera (Dage-MTI VE-1000; Sci-Tech) on predefined gain and black-level settings, and recorded for playback analysis using a video recorder. Four hours after LPS treatment, recordings were made of background fluorescence in the cremaster muscle at each of the excitation wavelengths, and these data were then subtracted from subsequent intensity readings. Mice then received 100 g of Alexa 488–labeled RMP-1 and 20 g of Alexa 594–labeled anti-KLH via intravenous injection. Antibodies were allowed to circulate for 5 minutes, and then the mouse was exsanguinated. P-selectin expression in the exsanguinated muscle was then quantitated exactly as previously described (29). Data were expressed as 1) millimeters of positive vessels per square-millimeter of tissue area and 2) the mean relative fluorescence intensity. Flow cytometric analysis of leukocyte P-selectin– binding activity. Whole blood was isolated from wild-type and MIF⫺/⫺ mice, and RBCs were lysed with ammonium chloride. Aliquots (100 l) of the resultant leukocyte suspension were incubated with 5 g of P-selectin IgG fusion protein (BD 3026 GREGORY ET AL Table 1. Leukocyte rolling flux, velocity, adhesion, and emigration in untreated versus LPS-treated wild-type mice 5 hours after injection* Mouse group Rolling flux, cells/minute Rolling velocity, m/second Adhesion, cells/100 m Emigration, cells/field Untreated (n ⫽ 6) LPS-treated (n ⫽ 6) 48.4 ⫾ 8.9 89.6 ⫾ 12.4† 63.9 ⫾ 7.2 36.6 ⫾ 11.9† 3.0 ⫾ 0.3 34.2 ⫾ 7.2† 11.7 ⫾ 2.7 35.5 ⫾ 6.3† * Values are the mean ⫾ SEM. LPS ⫽ lipopolysaccharide (10 ng). † P ⬍ 0.05 versus untreated wild-type mice. Biosciences) for 30 minutes at 37°C. Cells were then washed and treated for 25 minutes with fluorescein isothiocyanate– conjugated anti-human IgG (Chemicon Australia, Boronia, Victoria, Australia) that had been preincubated with 5% mouse serum. The percentage of cells displaying P-selectin– binding activity and the mean fluorescence intensity of positive cells were determined by analysis in a MoFlo flow cytometer (Cytomation, Fort Collins, CO). These parameters were assessed for granulocytic and mononuclear cell populations, which were identified using forward and side scatter properties. Identification of infiltrated leukocytes via immunohistochemistry. LPS-treated leukocytes present in cremaster muscles were identified using 3-layer immunoperoxidase staining as previously described (34). Following fixation in paraformaldehyde–lysine–periodate, cryostat sections (6 m) of cremaster muscles were prepared and stained for neutrophils, macrophages, and T cells using mAb RB6-8C5 (anti–Gr1), mAb anti–FA-11 (anti-CD68), and mAb KT3 (anti-CD3), respectively (34). The numbers of each of these cell types were assessed on a scale of 0–4⫹. Statistical analysis. Data are presented as the mean ⫾ SEM. Student’s t-tests were performed to compare experimental groups. P values less than 0.05 were considered significant. (Figure 1A). In addition, leukocyte rolling velocity 4 hours after LPS administration was significantly higher in MIF⫺/⫺ mice (mean ⫾ SEM 56.4 ⫾ 10.7 m/second) RESULTS Reduction of LPS-induced leukocyte recruitment in MIFⴚ/ⴚ mice. Since LPS is one of the best characterized stimuli of MIF release, we first examined leukocyte trafficking induced by this mediator. Local treatment of the cremaster muscle with LPS (10 ng for 4–5 hours) induced a characteristic pattern of increased leukocyte rolling flux, decreased rolling velocity, and increased adhesion and emigration in wild-type mice (Table 1), findings that are consistent with previous observations (35). LPS treatment was also associated with a 1.6-fold increase in MIF mRNA in the cremaster muscle of wild-type mice, as assessed by real-time PCR (average data from 4 mice) (data not shown). We then used this model to compare leukocyte trafficking in wild-type and MIF⫺/⫺ mice. In LPS-treated MIF⫺/⫺ mice, leukocyte rolling flux was reduced by ⬃50% relative to that in similarly treated wild-type mice 4–5 hours after LPS treatment Figure 1. Leukocyte rolling flux (A), adhesion (B), and emigration (C) in cremasteric postcapillary venules of lipopolysaccharide (LPS)–treated (10 ng) wild-type (n ⫽ 6) and MIF⫺/⫺ (n ⫽ 6) mice at 4, 4.5, and 5 hours after LPS treatment. ⴱ ⫽ P ⬍ 0.05 versus LPS-treated wild-type mice. LEUKOCYTE INTERACTIONS IN THE MICROCIRCULATION OF MIF⫺/⫺ MICE 3027 Table 2. Hemodynamic parameters and circulating leukocyte counts in LPS-treated and TNF␣-treated wild-type and MIF⫺/⫺ mice 4–5 hours after treatment* Shear rate, seconds⫺1 Venule diameter, m 4 hours 4.5 hours 5 hours Leukocyte count, ⫻10⫺6/ml 30 ⫾ 1.6 28 ⫾ 1.0 350 ⫾ 57 668 ⫾ 109† 425 ⫾ 110 613 ⫾ 104 449 ⫾ 119 601 ⫾ 105 5.8 ⫾ 1.3 7.2 ⫾ 0.5 28 ⫾ 1.0 29 ⫾ 1.0 526 ⫾ 37 688 ⫾ 51† 486 ⫾ 34 645 ⫾ 63† 470 ⫾ 60 634 ⫾ 62† 5.7 ⫾ 1.1 12.3 ⫾ 1.6† LPS-treated mice Wild-type (n ⫽ 6) MIF⫺/⫺ (n ⫽ 6) TNF␣-treated mice Wild-type (n ⫽ 6) MIF⫺/⫺ (n ⫽ 6) * Values are the mean ⫾ SEM. LPS ⫽ lipopolysaccharide (10 ng); TNF␣ ⫽ tumor necrosis factor ␣ (50 ng). † P ⬍ 0.05 versus wild-type mice receiving the same treatment. than in wild-type mice (26.3 ⫾ 3.6 m/second) (P ⬍ 0.012). Leukocyte adhesion was also significantly reduced in MIF⫺/⫺ mice at 4 and 5 hours (Figure 1B). Moreover, 5 hours after LPS treatment, the number of emigrated cells in MIF⫺/⫺ mice was significantly reduced relative to that in similarly treated wild-type mice (Figure 1C). These reductions were not due to differences in the number of circulating leukocytes, since this value did not differ significantly between the strains (Table 2). However, the normal reduction in microvascular shear rate induced by LPS was absent in MIF⫺/⫺ mice, such that 4 hours after LPS dosing, the shear rate was significantly elevated in MIF⫺/⫺ mice relative to that in wild-type mice (Table 2). This difference was no longer significant at 4.5 and 5 hours post-LPS. Immunohistochemical analysis of the cremaster muscle indicated that in wild-type mice, the infiltrate consisted predominantly of Gr-1⫹ cells (neutrophils) and FA-11⫹ cells (monocyte/macrophages), with occasional KT3⫹ cells (lymphocytes) (Table 3). Although total leukocyte numbers were lower in the cremaster muscles of MIF⫺/⫺ mice, the leukocyte subset proportions were similar in both groups. Reduction of exteriorization-induced leukocyte rolling in MIF ⴚ/ⴚ mice. To determine whether leukocyte–endothelial cell interactions were altered in MIF⫺/⫺ mice in the absence of exogenous mediators of inflammation, we compared interactions in untreated wild-type and MIF⫺/⫺ mice. Initially after tissue exteriorization, leukocyte rolling was significantly reduced in MIF⫺/⫺ mice relative to that in wild-type mice (Figure 2A). At the same time point, rolling velocity was significantly increased (P ⬍ 0.02) in MIF⫺/⫺ mice (65.8 ⫾ 9.4 m/second) compared with wild-type mice (38.3 ⫾ 6.6 m/second) (Figure 2B). However, 30 minutes after exteriorization, leukocyte rolling flux in wild-type mice was reduced by ⬃40% and leukocyte rolling velocity was increased, such that both parameters were no longer statistically significantly different from those in MIF⫺/⫺ mice. Leukocyte adhesion and emigration were consistently low in both groups and were not significantly different. Microvascular shear rate did not differ between the groups throughout the observation period (data not shown). Reduction in TNF␣-induced leukocyte rolling and adhesion in MIFⴚ/ⴚ mice. To determine if the reduction in interactions observed in MIF⫺/⫺ mice was restricted to LPS or whether they also occurred in response to other inflammatory mediators, we examined leukocyte recruitment induced by TNF␣. In wild-type mice, TNF␣ (50 ng) induced a significant reduction in Table 3. Immunohistochemical assessment of infiltrated leukocytes in cremaster muscles of wild-type and MIF⫺/⫺ mice 4 hours after lipopolysaccharide treatment* Wild-type mice Individual scores Group mean MIF⫺/⫺ mice Individual scores Group mean Gr-1 positive FA-11 positive KT3 positive 2⫹, 3⫹, 1⫹ 2 3⫹, 2⫹, 2⫹ 2.3 1⫹, 1⫹, 1⫹ 1 1⫹, 1⫹, 1⫹, 1⫹ 1 1⫹, 2⫹, 1⫹, 1⫹ 1.3 1⫹, 1⫹ 1 * The numbers of neutrophils (Gr-1 positive), monocyte/macrophages (FA-11 positive), and lymphocytes (KT3 positive) were assessed on a scale of 0–4⫹. 3028 GREGORY ET AL Figure 2. Leukocyte rolling flux (A) and rolling velocity (B) in cremasteric postcapillary venules of untreated wild-type (n ⫽ 6) and MIF⫺/⫺ (n ⫽ 6) mice at 0, 30, and 60 minutes after tissue exteriorization. ⴱ ⫽ P ⬍ 0.05 versus untreated wild-type mice. rolling velocity and significant increases in adhesion and emigration, as previously described (36). Four hours after TNF␣ treatment, leukocyte rolling flux did not differ significantly between wild-type and MIF⫺/⫺ mice; however, by 4.5 and 5 hours, leukocyte rolling flux in MIF⫺/⫺ mice was significantly reduced relative to that in wild-type mice (Figure 3A). In addition, the characteristic reduction in leukocyte rolling velocity induced by TNF␣ was significantly attenuated in MIF⫺/⫺ mice at all time points from 4 to 5 hours (at 4 hours, 17.2 ⫾ 2.6 m/second in wild-type mice versus 38.4 ⫾ 3.5 m/ second in MIF⫺/⫺ mice; P ⬍ 0.001). Levels of leukocyte adhesion were also significantly decreased in MIF⫺/⫺ mice relative to those in wild-type mice (Figure 3B). As was seen in LPS-treated mice, microvascular shear rates following TNF␣ treatment were significantly higher in MIF⫺/⫺ mice than in wild-type mice (Table 2). In contrast, TNF␣-induced leukocyte emigration did not differ significantly between the 2 groups at 5 hours after TNF␣ administration (Figure 3C). Figure 3. Leukocyte rolling flux (A), adhesion (B), and emigration (C) in cremasteric postcapillary venules of tumor necrosis factor ␣ (TNF␣)–treated (50 ng) wild-type (n ⫽ 6) and MIF⫺/⫺ (n ⫽ 6) mice at 4, 4.5, and 5 hours after initiation of TNF␣ treatment. ⴱ ⫽ P ⬍ 0.05 versus TNF␣-treated wild-type mice. Roles of endothelial selectins in LPS-induced leukocyte rolling. The decreased leukocyte rolling observed in LPS-treated MIF⫺/⫺ mice raised the possibility of altered usage of the molecules responsible for rolling under these conditions. Therefore, we next determined which endothelial adhesion molecules were responsible for rolling in LPS-treated wild-type mice. In LPS-treated wild-type mice, P-selectin blockade eliminated ⬎95% of the leukocyte rolling. Similarly, in MIF⫺/⫺ mice, anti–Pselectin treatment virtually abolished rolling (Figure 4). In contrast, blockade of either E-selectin or VCAM-1 in LPS-treated wild-type mice did not alter leukocyte roll- LEUKOCYTE INTERACTIONS IN THE MICROCIRCULATION OF MIF⫺/⫺ MICE Figure 4. Role and expression of P-selectin in lipopolysaccharide (LPS)–treated wild-type and MIF⫺/⫺ mice 4 hours after LPS (10 ng) treatment. A, Effects of P-selectin blockade on leukocyte rolling flux in cremasteric postcapillary venules of LPS-treated wild-type (n ⫽ 6) and MIF⫺/⫺ (n ⫽ 6) mice. ⴱ ⫽ P ⬍ 0.05 versus LPS-treated MIF⫺/⫺ mice. mAb ⫽ monoclonal antibody. B and C, LPS-induced P-selectin expression in cremasteric postcapillary venules in wild-type (WT; n ⫽ 7) and MIF⫺/⫺ (n ⫽ 7) mice. The average length of vessels positive for P-selectin expression is shown in B; the mean intensity of P-selectin expression within individual postcapillary venules is shown in C. ing flux or velocity significantly (data not shown), indicating that these adhesion molecules played no role in mediating rolling in this low-dose LPS model. Together, these findings indicate that in this model of LPS treatment, P-selectin is the predominant mediator of rolling in both wild-type and MIF⫺/⫺ mice and that the reduced leukocyte rolling observed in MIF⫺/⫺ mice is due to a reduction in P-selectin–dependent interactions. To determine if the reduction in P-selectin– dependent interactions in MIF⫺/⫺ mice was due to an alteration in endothelial P-selectin expression, we next 3029 compared LPS-induced P-selectin expression in wildtype and MIF⫺/⫺ mice, using fluorochrome-conjugated mAb. It was previously established that local LPS induces up-regulation of P-selectin in the cremaster muscle (35). Comparison of P-selectin expression in wildtype and MIF⫺/⫺ mice 4 hours after LPS treatment revealed that the number of vessels expressing P-selectin and the level of P-selectin expression within individual venules did not differ significantly between the 2 types of mice (Figures 4B and C). It is also conceivable that the reduced rolling reflected an altered ability of circulating leukocytes from MIF⫺/⫺ mice to bind P-selectin. Therefore, we assessed P-selectin binding activity in leukocytes from each mouse strain. Examination of both the granulocyte compartment and the mononuclear leukocytes revealed that there was no difference in P-selectin binding activity between leukocytes from wild-type and MIF⫺/⫺ mice (data not shown). MIF and rapid expression of P-selectin. In murine endothelial cells, mediators such as TNF␣ and LPS induce expression of P-selectin in a protein synthesis– dependent manner (37). However, mediators such as histamine are capable of inducing P-selectin expression via an alternate pathway, such that preformed P-selectin is rapidly (within 5 minutes) translocated to the endothelial surface (30). In the present study, the observation that the initial increase in leukocyte rolling seen in the cremaster muscle of naive wild-type mice immediately after exteriorization (known to be P-selectin–dependent ) was absent in MIF⫺/⫺ mice raised the possibility that MIF may promote this rapid translocation of P-selectin. Therefore, we compared histamine-induced leukocyte rolling in cremasteric postcapillary venules in wild-type and MIF⫺/⫺ mice (Figure 5). In wild-type mice, histamine (100 M) induced a rapid increase in leukocyte rolling, peaking 5 minutes after initial exposure, before partially abating to a level of ⬃80 cells/minute. At the peak of the rolling response, treatment with anti–P-selectin mAb eliminated rolling (data not shown), indicating that this response was mediated exclusively by P-selectin. Histamine-treated MIF⫺/⫺ mice showed a similar initial profile of response, with rolling peaking at the same time and the same level as in wild-type mice. However, in contrast to wild-type mice, in MIF⫺/⫺ mice, rolling rapidly returned to basal levels (⬃40 cells/minute), such that 20 minutes after initiation of histamine superfusion, rolling in MIF⫺/⫺ mice was significantly lower than that in wild-type mice. This significant difference persisted for the remainder of the 60-minute histamine treatment (Figure 5). 3030 Figure 5. Effect of histamine (100 M) on leukocyte rolling flux in cremasteric postcapillary venules of wild-type (n ⫽ 6) and MIF⫺/⫺ (n ⫽ 6) mice. Histamine was superfused over the cremaster muscle continuously from time 0 to 60 minutes. ⴱⴱ ⫽ P ⬍ 0.05 versus the –5-minute time point (before histamine treatment) for each group of mice; ⌿ ⫽ P ⬍ 0.05 versus wild-type mice at the corresponding time point. Reduction of synovial leukocyte trafficking in MIFⴚ/ⴚ mice. Given the proinflammatory role of MIF in experimental arthritis and human RA, we were also interested in examining by intravital microscopy the role of MIF in regulating leukocyte trafficking in the synovial microvasculature. In untreated mice, leukocyte rolling was detected in synovial venules, consistent with previous observations (31). Use of anti–P-selectin mAb revealed that this constitutive rolling was entirely dependent on P-selectin (data not shown). Furthermore, similar to the observations in the cremaster muscle, the level of leukocyte rolling flux observed in synovial postcapillary venules immediately upon exteriorization was significantly reduced in MIF⫺/⫺ mice (Figure 6). Minimal leukocyte adhesion (⬍2 cells/100 m) was observed in untreated mice of either strain. We then examined the effects of carrageenan. Figure 7 illustrates that carrageenan induced significant increases in leukocyte rolling and adhesion in wild-type mice. Comparison of carrageenan-induced rolling flux and adhesion in wild-type and MIF⫺/⫺ mice revealed that these interactions were significantly reduced in MIF⫺/⫺ mice (Figures 8A and C). In addition, the leukocyte count in synovial lavage fluid was significantly reduced in MIF⫺/⫺ mice compared with wild-type mice (Figure 8D), indicating that the absence of MIF also resulted in a reduction in leukocyte egress from the synovial microvasculature. In both strains of mice, carrageenan-induced leukocyte rolling in the synovial microvasculature was abolished with an anti–P-selectin mAb (Figure 8B). Together, these findings illustrate that GREGORY ET AL Figure 6. Leukocyte rolling flux in synovial postcapillary venules of untreated wild-type (n ⫽ 7) and MIF⫺/⫺ (n ⫽ 8) mice. ⴱ ⫽ P ⬍ 0.05 versus wild-type mice. the ability of MIF to promote leukocyte recruitment extends to the synovial microvasculature and that, as seen in the LPS model, the absence of MIF was associated with a reduction in P-selectin–mediated rolling. DISCUSSION There is a growing body of evidence that MIF contributes to the development of inflammatory responses induced by a wide range of stimuli (3,8,10,13). However, whether MIF regulates one of the most critical aspects of inflammatory responses, leukocyte recruitment, remains unknown. Therefore, in order to examine this issue, we used intravital microscopy to directly assess leukocyte–endothelial cell interactions in the microvasculature of wild-type and MIF⫺/⫺ mice. In the cremasteric microvasculature, the absence of MIF was associated with significant reductions in leukocyte rolling and adhesion induced by LPS and TNF␣, as well as LPSinduced leukocyte emigration. Similar findings were observed in the synovial microvasculature in response to carrageenan, demonstrating that this putative function of MIF is also active in the setting of joint inflammation. MIF⫺/⫺ mice also displayed reductions in the leukocyte rolling induced by mild inflammatory stimuli such as tissue exteriorization and histamine treatment. Together, these findings reveal a previously undescribed function of MIF in the regulation of leukocyte– endothelial cell interactions and leukocyte entry into inflamed tissues and demonstrate an additional mechanism whereby MIF may promote inflammatory responses. Data from previous studies have raised the possibility that MIF directly promotes leukocyte recruitment in inflammation. MIF immunoneutralization has LEUKOCYTE INTERACTIONS IN THE MICROCIRCULATION OF MIF⫺/⫺ MICE 3031 Figure 8. Leukocyte rolling flux (A and B) and adhesion (C) in synovial vessels and leukocyte numbers in the joint space as assessed by joint lavage (D) in carrageenan-treated wild-type (n ⫽ 6) and MIF⫺/⫺ (n ⫽ 6) mice. The effect of P-selectin blockade on synovial leukocyte rolling in separate groups of carrageenan-treated wild-type (n ⫽ 4) and MIF⫺/⫺ (n ⫽ 5) mice is shown in B. ⴱ ⫽ P ⬍ 0.05 versus wild-type mice. mAb ⫽ monoclonal antibody. Figure 7. Intravital microscopic images of synovial postcapillary venules (broken lines indicate margins) of wild-type mice, demonstrating interacting rhodamine 6G–labeled leukocytes. A, Low level of leukocyte–endothelial cell interactions in an untreated knee. B, Marked increase in the number of leukocytes interacting with the venular endothelium 75 minutes after injection of carrageenan into the joint. Arrows show adherent leukocytes, with the remainder undergoing rolling interactions. C, Rolling flux and D, adhesion of leukocytes in synovial vessels after carrageenan (Carrag) treatment in wild-type mice compared with saline-treated wild-type mice. ⴱ ⫽ P ⬍ 0.05 versus saline-treated wild-type mice. been shown to reduce leukocyte numbers in inflamed tissues in models of DTH, arthritis, colitis, and EAE and in the lung in response to endotoxemia (11,16,17,24,38). However, whether these reductions in leukocyte accumulation were due to an ability of MIF to promote the entry of leukocytes into these sites has not been directly assessed. Indeed, an alternative explanation for these observations is that MIF promotes the survival of leukocytes that are present in inflammatory lesions. This possibility is supported by the finding that MIF can inhibit p53-dependent apoptosis (20,39). In the present study, however, the potential for apoptosis to affect the number of leukocytes present in tissues was minimized by the use of acute models of leukocyte recruitment. Over the time courses studied, it is unlikely that the effects of MIF on apoptosis or proliferation could explain the observed differences in the numbers of recruited leukocytes. An alternative hypothesis supported by the present observations is that MIF promotes leukocyte recruitment via effects on leukocyte– endothelial cell interactions. One possible explanation for the reduced interactions in MIF⫺/⫺ mice in some of the models examined here relates to effects on microvascular shear rate. The propensity for a leukocyte to undergo interactions with the endothelial surface is determined by the balance between proadhesive forces and the dispersive effects of the flowing blood, as evaluated by the microvascular shear rate (40). In both the LPS-treated and the TNF␣treated mice, the microvascular shear rate was signifi- 3032 cantly higher in MIF⫺/⫺ mice at various points between 4 and 5 hours after treatment. This correlated with reduced rolling and adhesive interactions within postcapillary venules in each of these models. It is possible that one of the myriad effects of MIF is to interact with an as-yet-unidentified pathway that controls the microvascular shear rate. The resultant reduction in this parameter then enhances the ability of leukocytes in the mainstream of the blood flow to undergo interactions with the endothelial surface. However, the observation that in histamine-treated mice, leukocyte rolling is reduced in MIF⫺/⫺ mice in the absence of an alteration in the microvascular shear rate indicates that this is not the sole mechanism whereby MIF promotes leukocyte– endothelial cell interactions. The decrease in rolling and adhesive interactions in MIF⫺/⫺ mice is suggestive of decreases in adhesion molecule expression and/or function. There are few existing data on the ability of MIF to regulate adhesion molecule expression. MIF blockade has been observed to result in a reduction in ICAM-1 expression in a model of gastric inflammation and in a reduction in ICAM-1 and VCAM-1 expression in experimental glomerulonephritis, providing indirect evidence of a role for MIF in supporting adhesion molecule expression (41,42). Analysis of the direct effects of MIF has shown that MIF can induce small increases in the expression of ICAM-1 by keratinocytes and dendritic cells in vitro (43). Furthermore, recombinant human MIF has been observed to induce the expression of ICAM-1 by endothelial cells (44). These experiments raise the possibility that MIF can directly mediate the up-regulation of other endothelial adhesion molecules. Given that all of the leukocyte rolling in the LPS-treated mice was P-selectin–dependent, we examined the effect of the absence of MIF on P-selectin expression in vivo following LPS treatment. These experiments showed that the level of P-selectin expression in postcapillary venules did not differ between wild-type and MIF⫺/⫺ mice. Similarly, leukocytes from wild-type and MIF⫺/⫺ mice displayed comparable abilities to bind murine P-selectin, suggesting that alterations in leukocyte function were not responsible for these observations. These findings indicated that alterations in endothelial P-selectin expression or leukocyte selectin ligand function were not responsible for the reduced interactions observed in the present study. Another possible explanation for the reduction in leukocyte–endothelial cell interactions in MIF⫺/⫺ mice is the absence of the MIF-mediated counterregulation of the antiinflammatory effects of glucocorticoids. MIF is GREGORY ET AL unique among proinflammatory cytokines in being a homeostatic regulator of the effects of glucocorticoids (18). This has been illustrated both in vitro, where MIF has been shown to override the glucocorticoid-mediated inhibition of cytokine release by LPS-stimulated human monocytes, and in vivo, by the observation that MIF can reverse the antiinflammatory effects of glucocorticoids during murine antigen-induced arthritis (16,18). These studies have led to the hypothesis that MIF and glucocorticoids reciprocally regulate inflammatory responses. It is increasingly well accepted that constitutively expressed glucocorticoids serve to limit a number of inflammatory processes. This ability extends to the regulation of rolling and adhesion induced by inflammatory stimuli such as LPS and histamine (45,46). In light of these findings, it is plausible that the reduced leukocyte–endothelial cell interactions observed in MIF⫺/⫺ mice result from unopposed activity of glucocorticoids in these animals. It has previously been shown that MIF deficiency is associated with a reduction in macrophage expression of the cellular receptor for LPS, TLR-4 (19). Indeed, this has been raised as a plausible mechanism for the reduced susceptibility of MIF⫺/⫺ mice to endotoxemiainduced mortality. It is conceivable that this mechanism was also responsible for the reduction in LPS-induced leukocyte recruitment observed in the present study. This prompted us to examine models of leukocyte recruitment induced by the alternative stimuli, TNF␣ and carrageenan. In both the TNF␣ and carrageenan models, it was noteworthy that rolling and adhesion were reduced in the absence of MIF, which is consistent with the observations from LPS treatment. Similarly, in the carrageenan model, leukocyte emigration was reduced in MIF⫺/⫺ mice. Given that neither of these stimuli is thought to utilize TLR-4, this indicates that mechanisms distinct from a reduction in TLR-4 expression are responsible for the modifications in leukocyte–endothelial cell interactions in MIF⫺/⫺ mice. It is noteworthy that in response to TNF␣, leukocyte rolling does not differ between wild-type and MIF⫺/⫺ mice 4 hours after stimulation, although subsequently, rolling decreases more rapidly in MIF⫺/⫺ mice. It is conceivable that this stems from differential alterations in adhesion molecule expression within this time frame. Indeed, in response to some mediators, MIF may not be required for leukocyte rolling flux to reach peak levels. However, it may have an important function in preventing re-internalization of adhesion molecules from the endothelial surface, thereby prolonging expression. This contention is supported by the observations LEUKOCYTE INTERACTIONS IN THE MICROCIRCULATION OF MIF⫺/⫺ MICE from the histamine model, in which the peak level of P-selectin–dependent rolling did not differ between the 2 strains, but rolling dissipated significantly more rapidly in MIF⫺/⫺ mice. Moreover, in contrast to the LPS model, no difference in TNF␣-induced leukocyte emigration was detected. During inflammatory responses, adherent leukocytes respond to chemotactic cues that cause them to exit the vasculature and enter tissues. It is probable that the chemotactic mediators that promote leukocyte emigration in response to local LPS administration differ from those at work in response to TNF␣. The present data suggest that MIF promotes the chemotactic mechanisms induced by LPS, but not those invoked by TNF␣. We have previously reported elevated expression of MIF in serum, synovial fluid, and cultured synoviocytes from RA patients, and have described an association between disease activity in RA, synovial cellularity, and synovial MIF expression (8,9). In the present study, the absence of MIF is associated with a reduction in leukocyte–endothelial cell interactions in inflamed cremasteric and synovial venules. These findings illustrate a novel mechanism whereby MIF may promote inappropriate inflammatory responses such as those which occur in RA. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. ACKNOWLEDGMENTS The authors would like to acknowledge the kind assistance of Dr. Andrew Issekutz (Dalhousie University, Halifax, Nova Scotia, Canada) for providing the RMP-1 and RME-1 antibodies and Drs. Dietmar Vestweber and Britta Engelhardt (Max Planck Institut, Muenster, Germany) for providing the 6C7.1 antibody. 18. 19. 20. REFERENCES 1. Bloom BR, Bennett B. Mechanism of a reaction in vitro associated with delayed-type hypersensitivity. Science 1966;153:80–2. 2. David JR. Delayed hypersensitivity in vitro: its mediation by cell-free substances formed by lymphoid cell-antigen interactions. Proc Natl Acad Sci U S A 1966;56:72–7. 3. Donnelly SC, Haslett C, Reid PT, Grant IS, Wallace WA, Metz CN, et al. Regulatory role for macrophage migration inhibitory factor in acute respiratory distress syndrome. Nat Med 1997;3: 320–4. 4. Onodera S, Suzuki K, Matsuno T, Kaneda K, Takagi M, Nishihira J. Macrophage migration inhibitory factor induces phagocytosis of foreign particles by macrophages in autocrine and paracrine fashion. Immunology 1997;92:131–7. 5. Calandra T, Bernhagen J, Mitchell RA, Bucala R. The macrophage is an important and previously unrecognized source of macrophage migration inhibitory factor. J Exp Med 1994;179: 1895–902. 6. Liew FY. Regulation of nitric oxide synthesis in infectious and autoimmune diseases. Immunol Lett 1994;43:95–8. 7. Onodera S, Tanji H, Suzuki K, Kaneda K, Mizue Y, Sagawa A, et 21. 22. 23. 24. 25. 26. 3033 al. High expression of macrophage migration inhibitory factor in the synovial tissues of rheumatism joints. Cytokine 1998;11:163–7. Leech M, Metz C, Hall P, Hutchinson P, Gianis K, Smith M, et al. Macrophage migration inhibitory factor in rheumatoid arthritis: evidence of proinflammatory function and regulation by glucocorticoids. Arthritis Rheum 1999;42:1601–8. Morand EF, Leech M, Weedon H, Metz C, Bucala R, Smith MD. Macrophage migration inhibitory factor in rheumatoid arthritis: clinical correlations. Rheumatology (Oxford) 2002;41:558–62. Bernhagen J, Calandra T, Mitchell RA, Martin SB, Tracey KJ, Voelter W, et al. MIF is a pituitary-derived cytokine that potentiates lethal endotoxaemia. Nature 1993;365:756–9. Bernhagen J, Bacher M, Calandra T, Metz CN, Doty SB, Donnelly T, et al. An essential role for macrophage migration inhibitory factor in the tuberculin delayed-type hypersensitivity reaction. J Exp Med 1996;183:277–82. Mikulowska A, Metz CN, Bucala R, Holmdahl R. Macrophage migration inhibitory factor is involved in the pathogenesis of collagen type II-induced arthritis in mice. J Immunol 1997;158: 5514–7. Leech M, Metz C, Santos L, Peng T, Holdsworth SR, Bucala R, et al. Involvement of macrophage migration inhibitory factor in the evolution of rat adjuvant arthritis. Arthritis Rheum 1998;41:910–7. Bozza M, Satoskar AR, Lin G, Lu B, Humbles AA, Gerard C, et al. Targeted disruption of migration inhibitory factor gene reveals its critical role in sepsis. J Exp Med 1999;189:341–6. Calandra T, Echtenacher B, Roy DL, Pugin J, Metz CN, Hulther L, et al. Protection from septic shock by neutralization of macrophage inhibitory factor. Nat Med 2000;6:164–70. Santos L, Hall C, Metz CN, Bucala R, Morand EF. Role of macrophage migration inhibitory factor (MIF) in murine antigeninduced arthritis: interactions with glucocorticoids. Clin Exp Immunol 2001;123:309–14. Denkinger CM, Denkinger M, Kort JJ, Metz C, Forsthuber TG. In vivo blockade of macrophage migration inhibitory factor ameliorates acute experimental autoimmune encephalomyelitis by impairing the homing of encephalitogenic T cells to the central nervous system. J Immunol 2003;170:1274–82. Calandra T, Bernhagen J, Metz CN, Spiegel LA, Bacher M, Donnelly T, et al. MIF as a glucocorticoid-induced modulator of cytokine production. Nature 1995;377:68–71. Roger T, David JR, Glauser MP, Calandra T. MIF regulates innate immune responses through modulation of Toll-like receptor 4. Nature 2001;414:920–4. Mitchell RA, Liao H, Chesney J, Fingerle-Rowson G, Baugh J, David J, et al. Macrophage migration inhibitory factor (MIF) sustains macrophage proinflammatory function by inhibiting p53: regulatory role in the innate immune response. Proc Natl Acad Sci U S A 2002;99:345–50. Granger N, Kubes P. The microcirculation and inflammation: modulation of leukocyte-endothelial cell adhesion. J Leukoc Biol 1994;55:662–75. Vestweber D, Blanks JE. Mechanisms that regulate the function of the selectins and their ligands. Physiol Rev 1999;79:181–213. Springer TA. Traffic signals for lymphocyte recirculation and leukocyte emigration: the multistep paradigm. Cell 1994;76: 301–14. Makita H, Nishimura M, Miyamoto K, Nakano T, Tanino Y, Hirokawa J, et al. Effect of anti-macrophage migration inhibitory factor antibody on lipopolysaccharide-induced pulmonary neutrophil accumulation. Am J Respir Crit Care Med 1998;158:573–9. Kanwar S, Bullard DC, Hickey MJ, Smith CW, Beaudet AL, Wolitzky BA, et al. The association between ␣4-integrin, Pselectin, and E-selectin in an allergic model of inflammation. J Exp Med 1997;185:1077–87. Hickey MJ, Issekutz AC, Reinhardt PH, Fedorak RN, Kubes P. Endogenous interleukin-10 regulates hemodynamic parameters, 3034 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. leukocyte-endothelial cell interactions, and microvascular permeability during endotoxemia. Circ Res 1998;83:1124–31. Yang YH, Hutchinson P, Leech M, Holdsworth SR, Morand EF. Anti-inflammatory effect of lipocortin-1 in experimental arthritis. Inflammation 1997;21:583–96. Hickey MJ, Granger N, Kubes P. Molecular mechanisms underlying IL-4-induced leukocyte recruitment in vivo: a critical role for the ␣4 integrin. J Immunol 1999;163:3441–8. Norman MU, Van De Velde N, Timoshanko JR, Issekutz A, Hickey MJ. Overlapping roles of endothelial selectins and VCAM-1 in immune complex-induced leukocyte recruitment in the cremasteric microvasculature. Am J Pathol 2003;163: 1491–503. Kubes P, Kanwar S. Histamine induces leukocyte rolling in post-capillary venules: a P-selectin-mediated event. J Immunol 1994;152:3570–7. Veihelmann A, Szczesny G, Nolte D, Krombach F, Refior HJ, Messmer K. A novel model for the study of synovial microcirculation in the mouse knee joint in vivo. Res Exp Med (Berl) 1998;198:43–54. Hickey MJ, Bullard DC, Issekutz A, James WG. Leukocyteendothelial cell interactions are enhanced in dermal postcapillary venules of MRL/faslpr (lupus-prone) mice: roles of P- and Eselectin. J Immunol 2002;168:4728–36. Drummond AE, Dyson M, Thean E, Groome NP, Robertson DM, Findlay JK. Temporal and hormonal regulation of inhibin protein and subunit mRNA expression by post-natal and immature rat ovaries. J Endocrinol 2000;166:339–54. Huang XR, Holdsworth SR, Tipping PG. Evidence for delayedtype hypersensitivity mechanisms in glomerular crescent formation. Kidney Int 1994;46:69–78. Yipp BG, Andonegui G, Howlett CJ, Robbins SM, Hartung T, Ho M, et al. Profound differences in leukocyte-endothelial cell responses to lipopolysaccharide versus lipoteichoic acid. J Immunol 2002;168:4650–8. McCafferty DM, Kanwar S, Granger DN, Kubes P. E/P-selectindeficient mice: an optimal mutation for abrogating antigen but not GREGORY ET AL 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. tumor necrosis factor-␣-induced immune responses. Eur J Immunol 2000;30:2362–71. Weller A, Isenmann S, Vestweber D. Cloning of the mouse endothelial selectins: expression of both E- and P-selectin is inducible by tumor necrosis factor-␣. J Biol Chem 1992;267: 15176–83. De Jong YP, Abadia-Molina AC, Satoskar AR, Clarke K, Rietdijk ST, Fauboin WA, et al. Development of chronic colitis is dependent on the cytokine MIF. Nat Immunol 2001;2:1061–6. Leech M, Lacey D, Xue JR, Santos L, Hutchinson P, Wolvetang E, et al. Regulation of p53 by macrophage migration inhibitory factor in inflammatory arthritis. Arthritis Rheum 2003;48:1881–9. Perry MA, Granger DN. Role of CD11/CD18 in shear-rate dependent leukocyte-endothelial cell interactions in cat mesenteric venules. J Clin Invest 1991;87:1798–804. Huang XR, Chun Hui CW, Chen YX, Wong BC, Fung PC, Metz C, et al. Macrophage migration inhibitory factor is an important mediator in the pathogenesis of gastric inflammation in rats. Gastroenterology 2001;121:619–30. Lan HY, Bacher M, Yang N, Mu X, Mikolic-Paterson DJ, Metz C, et al. The pathogenic role of macrophage migration inhibitory factor in immunologically induced kidney disease in the rat. J Exp Med 1997;185:1455–65. Stavitsky AB, Xianli J. In vitro and in vivo regulation by macrophage migration inhibitory factor (MIF) of expression of MHC-II, costimulatory, adhesion, receptor, and cytokine molecules. Cell Immunol 2002;217:95–104. Lin SG, Yu XY, Chen YX, Huang XR, Metz C, Bucala R, et al. De novo expression of macrophage migration inhibitory factor in atherogenesis in rabbits. Circ Res 2000;87:1202–8. Davenpeck KL, Zagorski J, Schleimer RP, Bochner BS. Lipopolysaccharide-induced leukocyte rolling and adhesion in the rat mesenteric microcirculation: regulation by glucocorticoids and role of cytokines. J Immunol 1998;161:6861–70. Suzuki H, Schmid-Schonbein GW, Suematsu M, Delano FA, Forrest MJ, Miyasaka M, et al. Impaired leukocyte-endothelial cell interaction in spontaneously hypertensive rats. Hypertension 1994; 24:719–27.