MRL/faslpr mice are affected by a systemic autoimmune disease that results in leukocyte recruitment to a wide range of vascular beds, including the cerebral microvasculature. The mechanisms responsible for the leukocyte trafficking to the brain in these animals are not known. Therefore, the aim of this study was to directly examine the cerebral microvasculature in MRL/faslpr mice and determine the molecular mechanisms responsible for this leukocyte recruitment. Intravital microscopy was used to assess leukocyte-endothelial cell interactions (rolling, adhesion) in the pial microcirculation of MRL+/+ (control) and MRL/faslpr mice at 8, 12, and 16 wk of age. Leukocyte rolling and adhesion were rarely observed in MRL+/+ mice of any age. MRL/faslpr mice displayed similar results at 8 and 12 wk. However, at 16 wk, significant increases in leukocyte rolling and adhesion were observed in these mice. Histological analysis revealed that the interacting cells were exclusively mononuclear. Leukocyte rolling was reduced, but not eliminated in P-selectin−/−-MRL/faslpr mice. However, leukocyte adhesion was not reduced in these mice, indicating that P-selectin-dependent rolling was not required for leukocyte recruitment to the cerebral vasculature in this model of systemic inflammation. E-selectin blockade also had no effect on leukocyte rolling. In contrast, blockade of either the α4 integrin or VCAM-1 eliminated P-selectin-independent leukocyte rolling. α4 Integrin blockade also significantly inhibited leukocyte adhesion. These studies demonstrate that the systemic inflammatory response that affects MRL/faslpr mice results in leukocyte rolling and adhesion in the cerebral microcirculation, and that the α4 integrin/VCAM-1 pathway plays a central role in mediating these interactions.
Systemic lupus erythematosus (SLE)3 is a debilitating systemic autoimmune disease in which the microvasculature of a wide range of organs is targeted by an inappropriate inflammatory response (1). Some of the more damaging consequences of this disease stem from injury to the CNS. Indeed, patients with SLE-related neurologic complications can be affected by symptoms ranging from mild cognitive impairments to seizures and strokes. Pathological investigations of brains from SLE patients have revealed microvascular injury, proceeding in many cases to cortical microinfarcts (2, 3). However, the mechanisms behind the inflammatory response in the CNS of SLE patients remain unclear.
Many insights into the immunological basis of SLE have been made by examination of various mouse strains affected by systemic autoimmune disease. One of the most widely studied of these, the MRL/faslpr mouse, displays many features in common with SLE patients, including increased autoantibody production, immune complex deposition, systemic vasculitis, and glomerulonephritis (4). The cerebral vasculature is also affected in these mice, with mononuclear cell infiltrates present surrounding cerebral microvessels and evidence of extravascular accumulation of IgG and albumin, suggestive of a disruption of the blood-brain barrier (5). Furthermore, MRL/faslpr mice develop evidence of neurological dysfunction illustrated by defects in balance, learning ability, and other cognitive impairments, temporally correlated with the onset of leukocyte recruitment (6, 7, 8). These observations raise the possibility that leukocyte recruitment to the CNS is a key factor in the neurologic complications in MRL/faslpr mice. However, the mechanisms responsible for recruitment of these leukocytes to the cerebral microvasculature have not been investigated.
There is now extensive evidence that leukocyte recruitment to sites of inflammation involves a sequence of interactions between circulating leukocytes and endothelial cells (9). Initially, leukocytes must tether and roll along the endothelial surface, before undergoing adhesion and emigrating out of the vasculature. The tethering and rolling steps are mediated by members of the selectin family of adhesion molecules, and the α4 integrin expressed on specific leukocyte populations (10, 11, 12, 13, 14). Subsequent leukocyte adhesion is mediated by interaction of leukocyte integrins (β2 and β1) with their respective endothelial ligands, including ICAM-1 and VCAM-1 (9). Although this paradigm has been supported by repeated observations in tissues, such as the mesentery and skeletal muscle, there is a growing body of evidence that the cerebral microvasculature responds to inflammatory stimulation in a highly unique manner.
Direct observation of the cerebral (pial) microcirculation has shown that constitutive leukocyte rolling is almost entirely absent in cerebral microvessels, in marked contrast to most other organs (15, 16). In addition, local injection of proinflammatory agents such as fMLP and TNF-α, which readily induce leukocyte recruitment to peripheral microvascular beds, fails to induce significant leukocyte recruitment in the brain (17). Systemic activation with TNF-α can induce leukocyte rolling and adhesion within the cerebral microvasculature, although the molecular mechanisms used in this process are distinct from those observed in the periphery (15). Finally, recent experiments have indicated that leukocyte rolling is not required for recruitment of encephalitogenic T cell blasts to the CNS microcirculation (18). Given these observations, it is conceivable that the mechanisms of leukocyte recruitment to the CNS during a systemic autoimmune disease such as lupus may be highly divergent from those at work in peripheral organs exposed to the identical stimulus. Therefore, the aim of these experiments was to examine the cerebral microvasculature of lupus-prone MRL/faslpr mice to determine the mechanisms of the leukocyte recruitment that occurs in the CNS of these mice. These experiments revealed that leukocyte-endothelial cell interactions are increased in cerebral microvessels of MRL/faslpr mice, and that these interactions are critically dependent on the α4 integrin/VCAM-1 pathway.
Materials and Methods
Lupus-prone MRL-MpJ/faslpr (MRL/faslpr) mice and MRL-MpJ+/+ (MRL+/+) mice were supplied by The Jackson Laboratory (Bar Harbor, ME). (MRL/faslpr mice have recently been renamed MRL/Tnfrsf6lpr; however, we will use the more familiar nomenclature in describing these experiments.) P-selectin−/−-MRL/faslpr mice were generated by backcrossing a gene-targeted P-selectin mutation onto the MRL/faslpr strain background, as previously described (19). As controls for the lupus-prone MRL/faslpr mice, we used MRL+/+ mice. This mouse strain has displayed a susceptibility to autoimmune disease later in life. However, at the ages examined in this study, we have previously observed no signs of inflammation in the dermal vasculature (19). Mice were used at 8, 12, and 16 wk of age and weighed between 30 and 50 g.
Animals were anesthetized by i.p. injection of a cocktail of 10 mg/kg xylazine (Bayer Pharmaceuticals, Pymble, New South Wales, Australia) and 200 mg/kg ketamine hydrochloride (Caringbah, New South Wales, Australia). A catheter was inserted in the tail vein to administer anesthetic, fluorescent dyes, and Abs. The animal was placed on a thermo-controlled heating pad, regulating the core temperature to 37°C.
The cerebral microcirculation was then prepared for microscopy, as previously described (15). Briefly, a craniotomy was performed on the right parietal bone using a high speed drill (Fine Science Tools, North Vancouver, British Columbia, Canada). A superfusion chamber was held in place over the craniotomy by securing the incised scalp around the lower lip of the chamber with dental cement and Loctite 401 rapid adhesive (Loctite Australia, Caringbah, New South Wales, Australia). Before removal of the dura and exposure of the pial microvasculature, artificial cerebrospinal fluid (CSF) (ionic composition in mmol/L: NaCl, 132; KCl, 2.95; CaCl2, 1.71; MgSO4, 1.4; NaHCO3, 24.6; glucose, 3.71; urea, 6.7; pH 7.4) was continuously pumped through the superfusion chamber at 37°C to maintain the exposed brain. A pH and gas tension similar to that of normal CSF was replicated in the artificial equivalent by constant bubbling with 12% O2, 5% CO2, and 83% N2. Animals were examined for a maximum of 1 h.
To visualize leukocytes, animals were injected with 50 μl of 0.05% (i.v.) rhodamine 6G (Sigma-Aldrich, St. Louis, MO) immediately before microscopy. Rhodamine 6G at the dose used labels leukocytes and platelets, and has been shown to allow detection of the same number of rolling leukocytes as transmitted light, and has no effect on leukocyte kinetics (20, 21). It therefore allows for quantification of leukocyte rolling flux, leukocyte rolling velocity, and leukocyte adhesion via epifluorescence microscopy. Rhodamine 6G-associated fluorescence was visualized by epi-illumination at 510–560 nm, using a 590-nm emission filter (21, 22). To aid in visualization of the vasculature, 10 μl of 5% FITC/250-kDa dextran (Sigma-Aldrich) was administered i.v. as a plasma marker. This fluorochrome was visualized by epi-illumination at 450–490 nm, with a 520-nm emission filter.
The brain microvasculature was visualized using an intravital microscope (Axioplan 2 Imaging; Carl Zeiss, Carnegie, Victoria, Australia) with a ×40 water immersion objective lens (Achroplan ×40/0.80 NA; Carl Zeiss) and a ×10 eyepiece. A SIT video camera (Dage-MTI VE-1000; Sci Tech Pty., Preston South, Australia) was used to project the images onto a monitor (Sony PVM-20N5E; Carl Zeiss), and the images were recorded for playback analysis using a videocassette recorder (Panasonic NV-HS950; Klapp Electronics, Prahran, Victoria, Australia). One to four pial postcapillary venules (30–50 μm in diameter) were selected in each experiment, and to minimize variability, the same section of venule was observed throughout the experiment. Venular diameter and the number of rolling and adherent leukocytes were determined off-line during video playback analysis. Rolling leukocytes were defined as those cells moving at a velocity less than that of erythrocytes within a given vessel. Leukocyte rolling velocity was determined by measuring the time required for a leukocyte to roll along a 100-μm length of venule. Rolling velocity was determined for 20 leukocytes at each time interval. In some animals, however, less than 20 leukocytes were observed rolling in a vessel during the period of recording. In these animals, the velocity of each of the leukocytes observed to be rolling was measured. Leukocytes were considered adherent to the venular endothelium if they remained stationary for 30 s or longer. To account for variability of venular diameter, leukocyte adhesion was expressed as cells/mm2 of venular surface area, as shown previously (15).
RBC velocity was determined via analysis of the velocity of 1-μm-diameter fluorescent polystyrene microspheres (FluoSpheres, yellow/green; Molecular Probes, Eugene, OR) injected i.v. in 10 μl boluses (23). Beads were visualized via epifluorescence as for FITC dextran. Video sequences showing microspheres moving through postcapillary venules were digitized using an image analysis computer with an MVC-IC-PCI video capture card (Coreco Imaging, St. Laurent, Quebec, Canada) controlled by the Sequence Snap video acquisition software (Adept Electronic Solutions, Perth, Western Australia, Australia). Following calibration appropriate to the magnification under examination, RBC velocity was measured using Scion Image (Scion, Frederick, MD), as previously described (19). The mean velocity (VMEAN) of 10 randomly selected microspheres was determined, and venular wall shear rate (γ) was calculated based on the Newtonian definition: γ = 8(VMEAN/Dv) (24).
The Abs used in vivo in this study were RB40.34, a mAb against murine P-selectin (BD Biosciences, San Diego, CA; 20 μg/mouse); R1-2, a mAb against the murine α4 integrin (BD Biosciences; 75 μg/mouse); RME-1, a mAb against rat and mouse E-selectin (100 μg/mouse; generously provided by A. Issekutz, Dalhousie University, Halifax, Nova Scotia, Canada); and MK/2.7, a mAb against murine VCAM-1 (100 μg/mouse; R&D Systems, Minneapolis, MN). The doses of all function-blocking Abs used have been shown previously to be effective in specifically blocking their respective target molecules in vivo (12, 14, 19, 25).
Whole brains were fixed in formalin, and 7-μm sections were prepared and stained with H&E according to standard techniques. Profiles of pial postcapillary venules, corresponding to those viewed in vivo, were identified, and the presence of leukocytes in close apposition to the endothelium was determined. All venular profiles with at least one leukocyte closely apposed to the endothelium were defined as containing leukocytes interacting with the endothelium. These leukocytes were then classified as either granulocytic or mononuclear according to their nuclear morphology.
Flow cytometric analysis of α4 integrin expression
Samples (100 μl) of whole blood from MRL+/+ and MRL/faslpr mice underwent erythrocyte lysis and paraformaldehyde fixation using a Q-Prep Workstation (Beckman Coulter, Miami, FL). Samples were then incubated with R1-2 (0.5 μg/106 cells) for 25 min. Cells were washed, then incubated with FITC-conjugated sheep anti-rat IgG (Silenus, Melbourne, Victoria, Australia) (1/100, 25 min), following preincubation with 5% mouse serum to prevent nonspecific binding. Cells were subsequently washed and analyzed using a MoFlo flow cytometer (Cytomation, Fort Collins, CO).
Cerebral microvascular permeability
Microvascular permeability of the cerebral microvasculature was assessed utilizing a modification of a technique used previously in rats (26). Mice were anesthetized and maintained at 37°C with a heating blanket, as for the intravital microscopy experiments. The left femoral artery and the tail vein were catheterized for arterial blood sampling and delivery of molecular tracers and additional anesthetic, respectively. The pial microvasculature was accessed via a cranial window, as for microscopy experiments. Artificial CSF was continuously superfused across the exposed vasculature at 0.8 ml/min and collected afterward via an outflow port in the superfusion chamber. At the completion of the surgery, the tissue underwent a 30-min equilibration period to allow any bleeding from the dural vasculature to cease, before administration of the intravascular tracer. Permeability of the exposed vasculature was assessed by measuring the clearance of 70-kDa FITC dextran (Sigma-Aldrich) from the pial vessels into the superfused artificial CSF (26). At the start of the experimental period, FITC dextran was given as a bolus dose (1.25 mg/10 g of 5% solution, in 200 μl heparinized saline, i.v.). Subsequently, the artificial CSF was collected for the last minute of each 5-min period throughout the 30-min experimental period. Arterial blood samples (50 μl) were taken at 15 and 30 min, and 10 μl plasma samples were collected. The concentration of FITC-derived fluorescence in the CSF solution and plasma samples was measured on a 96-well plate fluorimeter (λex485 and λem538) (PolarStar Optima; BMG Labtechnologies, Mount Eliza, Victoria, Australia), and concentrations were determined by reference to a standard curve. FITC dextran clearance was determined by multiplying the ratio of CSF solution concentration to plasma concentration by the CSF flow rate.
For parameters such as leukocyte rolling flux, rolling velocity, and adhesion, within strain comparisons of mice at different ages were performed with one-way ANOVA. Comparisons between the two mouse strains were performed using Student’s t tests. Comparisons of rolling and adhesion before and after administration of mAbs were performed using paired t tests. A value of p < 0.05 was deemed significant.
Leukocyte-endothelial cell interactions are increased in the cerebral microvasculature of MRL/faslpr mice
In initial experiments, we examined the pial microcirculation of MRL+/+ and MRL/faslpr mice at 8, 12, and 16 wk. Previously, we had observed that these time points were before the onset of severe disease in MRL/faslpr mice, and animals remained sufficiently healthy to undergo the anesthesia and microscopy procedures. In pial postcapillary venules of MRL+/+ mice at all ages examined, leukocyte rolling was rarely observed (Fig. 1). This is in concert with previous observations of the pial circulation in uninflamed, wild-type mice (15). In MRL/faslpr mice at 8 and 12 wk, leukocyte rolling was similarly infrequent (1–2 cells/min). However, in 16-wk-old mice, leukocyte rolling was significantly increased to an average of 20 cells/min (Fig. 1). Leukocyte adhesion followed a similar pattern. Minimal adhesion was observed in MRL+/+ mice at any age (Fig. 2). In MRL/faslpr mice at 8 and 12 wk, leukocyte adhesion was at comparable levels to that observed in MRL+/+ mice. However, at 16 wk, leukocyte adhesion was significantly increased above levels in comparably aged MRL+/+ mice, and above levels in MRL/faslpr mice at 8 and 12 wk (Fig. 2). These differences in leukocyte-endothelial cell interactions in 16-wk-old mice were not due to divergences in venular diameters or hydrodynamic shear rates, as these did not differ significantly between the two strains (Table I). Furthermore, we have previously reported that the number of circulating white blood cells did not significantly differ between the two strains of mice at the ages under examination (19).
|.||MRL+/+ .||MRL/faslpr .|
|Venular diameter (μm)||42 ± 4||42 ± 3|
|(n = 5)||(n = 7)|
|Shear rate (s−1)||167 ± 43||241 ± 72|
|(n = 3)||(n = 6)|
|.||MRL+/+ .||MRL/faslpr .|
|Venular diameter (μm)||42 ± 4||42 ± 3|
|(n = 5)||(n = 7)|
|Shear rate (s−1)||167 ± 43||241 ± 72|
|(n = 3)||(n = 6)|
Data are shown as mean ± SEM of n mice, as indicated for each group.
The clear difference between the two strains in the total number of leukocytes interacting with the endothelial surface is illustrated by the images of the pial microcirculation shown in Fig. 3. Although the levels of adherent cells in 16-wk MRL/faslpr mice are low compared with some other types of inflammatory response, they are comparable to the increased levels of adhesion observed previously in the dermal microcirculation of 16-wk MRL/faslpr mice (19). Furthermore, the levels of adhesion observed in this model of chronic, systemic inflammation reach ∼50% of that achieved in this microvasculature by systemic treatment with high doses of the proinflammatory cytokine, TNF-α (15). Histological analysis of pial vessels in brains from each strain also showed that in 16-wk MRL/faslpr mice, the number of venules containing leukocytes apposed to the endothelial surface was significantly increased relative to MRL+/+ mice (Table II). Morphological analysis of these leukocytes indicated they were almost exclusively mononuclear. Finally, as previous studies of the cerebral microvasculature have observed that platelets readily undergo interactions with this vascular bed, and indeed are capable of promoting interactions between leukocytes and the endothelium, we assessed platelet recruitment in 16-wk MRL/faslpr mice (15). However, very few platelets were observed to undergo interactions in the cerebral microvasculature of these animals, apart from occasional transient tethering interactions in some mice.
|.||Mean No. of Vascular Profiles Examined/Mouse .||Percentage of Profiles with Leukocytes .|
|MRL+/+||25||4.2 ± 0.9%|
|MRL/faslpr||48||19.4 ± 6.6%∗|
|.||Mean No. of Vascular Profiles Examined/Mouse .||Percentage of Profiles with Leukocytes .|
|MRL+/+||25||4.2 ± 0.9%|
|MRL/faslpr||48||19.4 ± 6.6%∗|
Shown are the number of vascular profiles assessed, and the percentage exhibiting leukocytes apposed to the endothelial surface. Data are shown as mean ± SEM of observations from six mice per group. ∗, Denotes p < 0.05 vs MRL+/+ mice.
Role of endothelial selectins in mediating increased rolling in cerebral microvasculature of MRL/faslpr mice
We next attempted to determine the roles of P- and E-selectin in mediating the increased leukocyte rolling observed in 16-wk MRL/faslpr mice. Two approaches were used to examine the role of P-selectin. First, the effect of function-blocking mAbs against P-selectin in wild-type MRL/faslpr mice was determined (Fig. 4,A). In mice with the most markedly elevated leukocyte rolling (20–30 cells/min), P-selectin blockade reduced leukocyte rolling to ∼4 cells/min. However, in 50% of mice examined, leukocyte rolling was in the range 4–10 cells/min. In these animals, anti-P-selectin had no appreciable effect. To further delineate the role of P-selectin, we examined P-selectin-deficient MRL/faslpr mice at 16 wk. The level of leukocyte rolling in these mice was consistently observed to be ∼4 cells/min (Fig. 4,B), comparable to that seen in wild-type MRL/faslpr mice after treatment with anti-P-selectin mAb. Analysis of leukocyte rolling velocities in wild-type MRL/faslpr mice showed that before P-selectin blockade, leukocytes were rolling at a wide range of velocities, from <10 μm/s to >100 μm/s (mean = 40.5 μm/s) (Fig. 5). However, following P-selectin blockade, the remaining leukocytes rolled more slowly, with no rolling observed above a velocity of 60 μm/s (mean = 28.0 μm/s). Similarly, in P-selectin−/−-MRL/faslpr mice, leukocyte rolling was not observed at velocities greater than 50 μm/s (mean = 6.6 μm/s) (Fig. 5). These observations suggested that the main function of P-selectin in the cerebral microvasculature of MRL/faslpr mice was to support the rolling of leukocytes at velocities above 50–60 μm/s. However, alternative molecules appear to be more important in supporting rolling at lower velocities.
Finally, the importance of P-selectin-mediated leukocyte rolling for subsequent leukocyte adhesion was assessed by examining adhesion in wild-type and P-selectin−/−-MRL/faslpr mice. Levels of leukocyte adhesion in pial venules in P-selectin−/−-MRL/faslpr mice did not differ from those in wild-type MRL/faslpr mice, indicating that P-selectin-dependent rolling was not critical in allowing cells to progress to adhesion (Fig. 6).
To examine the role of E-selectin in mediating the rolling observed in the absence of P-selectin, 16-wk P-selectin−/−-MRL/faslpr mice were treated with anti-E-selectin mAb. This treatment had no effect on leukocyte rolling flux (Fig. 7). Given that E-selectin has previously been shown to be important in reducing leukocyte rolling velocity in inflamed microvessels (27), velocity was also assessed in these animals. Mean rolling velocity did not alter following E-selectin blockade (27.2 ± 4.2 μm/s pre-mAb vs 25.9 ± 3.1 μm/s post-mAb; mean ± SEM, n = 5). Together these observations indicated that E-selectin played no role in these interactions.
Critical role of α4 integrin in leukocyte-endothelial cell interactions in cerebral microvasculature of lupus-prone mice
As the α4 integrin has been shown to mediate leukocyte rolling in other models of chronic inflammatory disease (13), we next examined the effect of α4 integrin blockade. α4 Integrin blockade in 16-wk wild-type MRL/faslpr mice significantly reduced leukocyte rolling, and also reduced leukocyte adhesion to basal levels within a 20-min period of mAb treatment (Fig. 8, A and B). Residual leukocyte rolling was observed in two of six animals. This rolling displayed the elevated rolling velocity characteristic of P-selectin-dependent rolling observed earlier, and was eliminated by P-selectin blockade (data not shown). The effect of α4 integrin blockade was more striking in 16-wk P-selectin−/−-MRL/faslpr mice. In these animals, treatment with R1-2 eliminated all leukocyte rolling (Fig. 8,C). In addition, R1-2 significantly reduced adhesion, as seen in wild-type MRL/faslpr mice (Fig. 8 D). These observations indicate that the α4 integrin is essential for P-selectin-independent rolling in this model, and may also directly mediate leukocyte adhesion.
To determine whether alterations in α4 integrin expression by circulating leukocytes contributed to the observed increase in α4 integrin-dependent interactions, we measured α4 integrin expression in leukocytes isolated from 16-wk MRL/faslpr and MRL+/+ mice. These experiments revealed no quantitative differences in α4 integrin expression by circulating leukocytes from the two strains of mice (data not shown).
VCAM-1 mediates leukocyte-endothelial cell interactions in the cerebral microvasculature of lupus-prone mice
As it has been previously reported that the potential α4 integrin ligand VCAM-1 is expressed at elevated levels in the brains of MRL/faslpr mice at 14 wk of age (28), we examined the role of VCAM-1 in cerebral leukocyte trafficking in these mice. Treatment of 16-wk MRL/faslpr mice with the anti-VCAM-1 mAb MK/2.7 reduced leukocyte rolling in most of the mice examined (Fig. 9,A). Moreover, all VCAM-1-independent residual rolling had a high rolling velocity (∼80 μm/s) characteristic of the P-selectin-dependent rolling previously observed in MRL/faslpr mice after treatment with an α4 integrin mAb. This residual rolling was abolished by P-selectin blockade (Fig. 9,A). Examination of leukocyte adhesion revealed a similar pattern in that adhesion was almost completely eliminated in four of six mice examined (Fig. 9 B). Finally, VCAM-1 blockade also profoundly reduced rolling in P-selectin−/−-MRL/faslpr mice (data not shown), similar to the effect of α4 integrin blockade in these animals. These data suggest that, as observed for the α4 integrin, VCAM-1 plays a critical role in mediating cerebral leukocyte trafficking in lupus-prone mice.
Cerebral microvascular permeability in MRL/faslpr mice
To determine whether the leukocyte recruitment observed in MRL/faslpr mice was associated with a disruption in the blood-brain barrier, we assessed cerebral microvascular permeability in 16-wk MRL/faslpr mice, and compared it with that in similarly aged MRL+/+ mice. No difference in leakage of 70-kDa FITC dextran was observed between the two strains of mice (Fig. 10). Similarly, cerebral microvascular permeability in P-selectin−/−-MRL/faslpr mice was found to be indistinguishable from that in MRL+/+ and MRL/faslpr mice.
There is a growing body of evidence that cerebral SLE is associated with microvascular injury and dysfunction. Despite this, little is known regarding the mechanisms of the inflammatory response that affects the cerebral microvasculature, and specifically the molecular basis for recruitment of leukocytes to this vascular bed, either in SLE patients, or in animal models of SLE. In the present study, we have investigated the cerebral microvasculature of lupus-prone MRL/faslpr mice and observed an increase in leukocyte-endothelial cell interactions at 16 wk of age. Histological analysis revealed that the vast majority of these interacting leukocytes were mononuclear. This observation is in concert with previous histological analysis of the brains of MRL/faslpr mice, in which extensive perivascular mononuclear leukocyte infiltration, consisting predominantly of CD4+ T cells, was observed in 22-wk-old mice (5). In the present study, we went on to determine the adhesion molecule pathways functioning in the cerebral microvasculature of MRL/faslpr mice. These studies revealed that the α4 integrin/VCAM-1 pathway was critical in mediating both rolling and adhesion in these animals. In contrast, experiments in both wild-type MRL/faslpr mice and P-selectin−/−-MRL/faslpr mice showed that P-selectin mediated some rolling in cerebral microvessels, but this rolling was mainly of a high velocity that was not critical in allowing leukocytes to ultimately adhere. These observations indicate that in this spontaneous model of systemic inflammatory disease, the α4 integrin/VCAM-1 pathway plays a key role in allowing recruitment of mononuclear leukocytes to the cerebral microvasculature.
We have previously documented a progressive increase in leukocyte-endothelial cell interactions in the dermal microcirculation of MRL/faslpr mice, involving both granulocytic and mononuclear cells (19). The enhanced rolling interactions in dermal microvessels were mediated predominantly via P-selectin, with E-selectin playing an additional minor role. These findings were associated with alterations in P-selectin glycoprotein ligand-1 expression by circulating cells, as well as changes in P-selectin expression in the dermal microvasculature. In contrast, in the present study, examination of the cerebral microcirculation in the same population of mice revealed that enhanced leukocyte recruitment in the cerebral microvasculature was mediated by the α4 integrin/VCAM-1 pathway, independently of P- and E-selectin. This divergence is not unexpected, in that several studies have shown that the molecular mechanisms of leukocyte recruitment vary between individual tissues, even in response to an identical inflammatory stimulus (29, 30, 31). Moreover, it has previously been demonstrated that expression of VCAM-1 is significantly increased in the brains of MRLfaslpr mice at 14 wk (28). The timing of this increase in VCAM-1 expression correlates well with the observed increase in VCAM-1-dependent interactions observed in the present study. Together, these findings suggest that the increase in VCAM-1 expression in the cerebral microcirculation is of key importance in mediating the observed increase in cerebral leukocyte trafficking in MRL/faslpr mice. It is noteworthy that despite previous evidence of alterations in P-selectin glycoprotein ligand-1 expression by circulating leukocytes in these mice, the minimal role of the endothelial selectins indicates that alterations in selectin ligand expression were without effect on leukocyte recruitment to the cerebral vasculature. Clearly, in this model of systemic inflammatory disease, changes in selectin ligand expression by circulating cells alone are insufficient to mediate an increase in leukocyte recruitment to this vascular bed. Additional alterations in adhesion molecule expression in the local microvasculature must also be required.
Although there are few other studies that have used intravital microscopy to examine the microvasculature in animals affected with a chronic systemic inflammatory disease, some novel observations of aberrant leukocyte trafficking have been made in rats with chronic adjuvant-induced vasculitis. Johnston et al. (13, 32) examined leukocyte trafficking in mesenteric postcapillary venules in rats following immunization with CFA, and noted that leukocyte-endothelial cell interactions increased markedly in the days following immunization. The molecular mechanisms responsible for these increased interactions were quite distinct from those at work under acute inflammatory conditions. As in the present study, it was found that the α4 integrin was of key importance in mediating the increases in both rolling and adhesion. Given that the α4 integrin is most highly expressed on mononuclear leukocytes and eosinophils, but not neutrophils, this suggested that these interacting cells were predominantly mononuclear leukocytes. This contention was supported by the observation that rolling and adhesion were not diminished by neutrophil depletion strategies (13). Similarly, in the present study, a role for the α4 integrin/VCAM-1 pathway was observed concurrent with an exclusively mononuclear infiltrate. This is consistent with a more prominent role for mononuclear leukocytes in chronic inflammatory responses.
Previous analysis of the cerebral microvasculature has shown that it responds to inflammatory stimulation in a highly unique manner. Constitutive leukocyte rolling is rarely detectable in the pial circulation, and intracerebral injection of acute inflammatory agents does not induce leukocyte recruitment to the CNS parenchyma (15, 16, 17). Nevertheless, the cerebral microvasculature is not entirely refractory to inflammatory stimuli. Systemic activation with TNF-α induces leukocyte rolling and adhesion in the pial circulation, via a highly unusual mechanism involving nonoverlapping roles for P-selectin, E-selectin, and platelets (15). Also, recruitment of encephalitogenic T cell blasts to the uninflamed microcirculation of the spinal cord has been observed to bypass the normally critical process of rolling, with these leukocytes capable of undergoing immediate capture and arrest on the endothelial surface in both capillaries and postcapillary venules (18). Interestingly, as seen in CNS microvessels in MRL/faslpr mice, in the latter study, the α4 integrin/VCAM-1 pathway was solely responsible for recruitment of encephalitogenic T cell blasts (18). The observation of the α4 integrin/VCAM-1 pathway mediating leukocyte recruitment independently of selectin function in these two studies further emphasizes the distinctive nature of recruitment to the cerebral vascular bed. This pathway has only been observed to operate rarely, in unique situations such as rolling of hemopoietic progenitor cells in bone marrow microvessels (25). However, our data show a further distinction from the observations of Vajkoczy et al. (18), in that we did not observe immediate capture of leukocytes in the cerebral microvasculature of MRL/faslpr mice. All recruited leukocytes were observed to undergo rolling in postcapillary venules. These findings further emphasize that even under the unique conditions of the cerebral microvasculature, different inflammatory states invoke alternative recruitment mechanisms.
The minimal role of P-selectin in this model may be considered somewhat surprising in light of previous examinations of cerebral inflammation. P-selectin blockade has been observed to reduce brain injury induced by permanent middle cerebral artery occlusion (33). Furthermore, direct observation of the cerebral microcirculation has revealed a critical role for P-selectin in short-term models, such as exposure to nicotine, systemic TNF-α, and LPS (15, 34, 35). Also, in a recent study of the experimental autoimmune encephalomyelitis model of CNS inflammation, leukocyte trafficking during active disease was shown to be highly dependent on P-selectin-mediated initiation of rolling (36). It is interesting to note that this study also revealed a role for the α4 integrin in mediating leukocyte rolling and adhesion, but in contrast to the present study, these α4 integrin-mediated interactions were critically dependent on initiation of rolling via P-selectin. There are several potential explanations for this difference. One possibility is that distinct leukocyte populations are being recruited in each of these studies. Indeed, neutrophils were the principal type of leukocyte recruited following systemic treatment with TNF-α, in contrast to the mononuclear leukocytes observed in the present study (15). Alternatively, the level of P-selectin expression in the brains of MRL/faslpr mice may not be elevated to the same extent as has been demonstrated in several of these models, including the TNF-α, LPS, and experimental autoimmune encephalomyelitis studies (15, 35, 36). It is noteworthy that measurement of P-selectin expression in various organs of MRL/faslpr mice has shown no increase from basal levels, despite the presence of ongoing inflammation in these animals (37). This suggests that the chronic inflammatory conditions present in these mice are not conducive to increased expression of P-selectin.
It has previously been reported that MRL/faslpr mice display evidence of enhanced macromolecular leakage across the blood-brain barrier (5). This was based on immunohistochemical analysis that revealed that IgG and IgM accumulated in extravascular areas in the brains of MRL/faslpr mice, with the amount of extravascular Ig increasing progressively from 8 to 26 wk. However, this observation was associated with a concomitant 5-fold increase in serum Ig. It is possible that the presence of these high levels of Ab in the serum may have falsely exaggerated the level of permeability assessed using this approach. Moreover, significant numbers of lymphocytes were also detectable in the brain parenchyma, where they may have acted as cellular sources of Ab behind the blood-brain barrier. In the present experiments, we assessed cerebral microvascular permeability using an in vivo technique that provides an instantaneous measurement of macromolecular leakage that is unaffected by the level of serum Ig or Ab production within the brain. Using this approach, we observed no difference between MRL+/+ and MRL/faslpr mice at 16 wk of age, despite the existence of significant levels of leukocyte adhesion in the latter group. This suggests that the level of inflammatory insult to the brain in MRL/faslpr mice at 16 wk is insufficient to compromise the function of the blood-brain barrier. However, this does not exclude the possibility that increases in permeability do occur in older mice, as the inflammatory response progresses.
Treatment of MRL/faslpr mice with immunosuppressive agents such as cyclophosphamide reduces both leukocyte recruitment into the brain and behavioral deficits (38). This suggests that attenuation of leukocyte recruitment to the brain may also be beneficial in SLE patients. The results of the present study illustrate that the α4 integrin/VCAM-1 pathway is of key importance in mediating leukocyte recruitment to the CNS in MRL/faslpr mice. This raises the possibility that this specific molecular pathway may also be of relevance in mediating CNS leukocyte recruitment in SLE patients. Clearly, further examination of the recruitment mechanisms at work in the CNS of lupus patients may determine whether these molecules present a relevant therapeutic target in cerebral lupus.
We acknowledge the generous assistance of Dr. Andrew Issekutz (Dalhousie University) for provision of anti-E-selectin mAb.
Funding for this study was provided by the National Health and Medical Research Council (NHMRC, Australia; Project Grant 166902). M.J.H. is an NHMRC R.D. Wright Fellow.
Abbreviations used in this paper: SLE, systemic lupus erythematosus; CSF, cerebrospinal fluid.