Microvascular Thrombosis and CD40/CD40L Signalling

Summary

Background

Although inflammation and thrombosis are now recognized to be interdependent processes that activate and perpetuate each other, the signalling molecules that link these two processes remain poorly understood.

Objectives

The objective of this study was to assess the contribution of the CD40/CD40L signalling system to the enhanced microvascular thrombosis that accompanies two distinct experimental models of inflammation, i.e., endotoxemia (LPS) and dextran sulphate sodium (DSS)-induced colitis.

Methods

Thrombosis was induced in cerebral (LPS model) and cremaster muscle (DSS model) arterioles and venules of wild type (WT) mice and mice deficient in either CD40 (CD40-/-) or CD40L (CD40L-/-) using the light/dye (photoactivation) method.

Results and conclusions

A comparison of thrombus formation between WT and mutant mice revealed a role for CD40 and/or CD40L in the inflammation-enhanced thrombosis responses in both the cerebral and muscle vasculatures. However, the relative contributions of CD40 and its ligand to thrombus formation differed between vascular beds (brain vs. muscle) and vessel types (arterioles vs. venules). The protective effect of CD40L deficiency in cerebral arterioles exposed to LPS was significantly blunted by administration of soluble CD40L. These findings implicate CD40 and its ligand in the enhanced thrombus formation that is associated with acute and chronic inflammation.

Introduction

While the critical role of platelets in hemostasis is now well recognized, there is mounting evidence that these blood cells also serve as both effector and target cells in a variety of acute and chronic inflammatory diseases. The effector function of platelets results from the capacity of these cells to liberate a broad range of mediators, including autacoids and cytokines, which can activate leukocytes and/or vascular endothelial cells. Platelet activation, on the other hand, can be elicited by substances released from inflammatory and endothelial cells. A consequence of these reciprocal interactions between platelets and inflammatory cells is an increased tendency for thrombus development during inflammation. The prothrombotic phenotype assumed by inflamed tissue also reflects an interdependency of coagulation and inflammation, with inflammatory mediators (e.g., cytokines) eliciting the activation of coagulation factors (e.g., tissue factor) and the generation of coagulation factors (e.g., thrombin) that induce an inflammatory phenotype in the vasculature (e.g., increased adhesion molecule expression) (1).

Both the arterial and venous segments of the vascular tree are vulnerable to thrombus formation. This is evident in both large (arteries and veins) and microscopic (arterioles and venules) blood vessels. Accelerated thrombus formation in the microvasculature has been demonstrated in the different animal models of inflammation, with some models preferentially enhancing thrombus formation in venules (e.g., endotoxemia), while others (e.g., hypercholesterolemia, colonic inflammation) exert the prothrombotic effect predominantly on arterioles (2,3). Mechanistic studies have revealed a role for different endogenous pro- and anticoagulant factors in inflammation-enhanced microvascular thrombosis. For example, an impaired protein C pathway has been implicated in the enhanced arteriolar thrombosis that accompanies experimental colitis (4), while an increased expression of von Willebrand factor (vWf) appears to underlie the enhanced venular thrombosis associated with endotoxemia (3).

Although considerable attention has been focussed on defining the signalling pathways that link inflammation to enhanced thrombus formation, much remains to be learned. Toll-like receptor-4 (TLR-4)-dependent signalling in endothelial cells has been implicated in endotoxin (LPS) enhanced venular thrombosis (6). The CD40/CD40L signalling pathway has received considerable attention as a contributor to atherothrombosis (7,8) and it has been shown to participate in thrombus formation and stability in non-inflamed mesenteric arterioles following ferric chloride-induced collagen exposure (9). While it remains unclear whether this signalling pathway is involved in inflammation-enhanced thrombus formation in the microvasculature, the ubiquitous expression of the receptor (CD40) or its ligand (CD40L) on cells that participate in both inflammation and hemostasis suggests that the CD40/CD40L system may link the two processes. Further support for this possibility is provided by clinical evidence that soluble CD40L (sCD40L), which is shed from the surface of activated platelets, is a prognostic marker of thrombotic risk in cardiovascular disease as well as inflammatory bowel disease (10).

The overall objective of this study was to evaluate the role of the CD40/CD40L system in inflammation-enhanced microvascular thrombosis. CD40-/- and CD40L-/- mice as well as soluble CD40L were used to assess the contribution of each component of the CD40/CD40L dyad. The studies were carried out in two distinct models of inflammation-enhanced thrombosis (LPS administration and dextran sodium sulphate-induced colonic inflammation) that have been previously reported to preferentially target either arterioles (DSS) or venules (LPS). Our findings indicate that the CD40/CD40L system contributes significantly to the enhanced microvascular thrombosis associated with both models of inflammation.

Methods

Animals

Male wild-type mice (WT: C57BL/6), B6.129S2-Tnfsf5tm1Imx (CD40L^−/−) mice, and B6.129P2-Tnfrsf5tm1Kik (CD40^−/−) mice, 25-35g, were obtained from Jackson Laboratories (Bar Harbor, Me; gene-targeted mice were on a C57BL/6 background). The mice were maintained on standard laboratory chow and fed ad libitum with normal drinking water either until immediately before the experiments (brain protocols) or until placement on the DSS regimen. All experimental procedures involving the use of animals were reviewed and approved by the Institutional Animal Care and Use Committee of LSU Health Sciences Center and performed according to the criteria outlined by the National Institutes of Health.

Dextran sodium sulphate (DSS)-induced colitis

As described previously (11), experimental colitis in mice (WT, CD40L^−/− and CD40^−/−) was induced by feeding 3% DSS (molecular weight, 40kDa; ICN Biomedicals, Aurora, OH) in millipore-filtered water, ad libitum, beginning on day 0 and ending on day 6. This DSS regimen is not associated with mortality. Control WT mice received filtered water alone (no DSS). Body weight, fecal status, presence of occult blood in stools, and perianal bleeding were evaluated daily to monitor the progression of colitis. Occult blood was detected using guaiac paper (Colo-Screen; Helena Laboratories) (11). A disease activity index, a measure of disease severity ranging between 0 and 4, was calculated from data collected on stool consistency, presence or absence of fecal blood, and weight loss as previously described (12). The disease activity index was monitored to ensure that DSS treatment resulted in clinical responses that are consistent with disease activity.

Endotoxin (LPS) administration

An acute prothrombotic state was induced by intraperitoneal administration of LPS into WT, CD40L^−/−and CD40^−/− mice. The LPS, Escherichia coli serotype 0111:B4 (purified initially by phenol extraction and purified further by ion exchange chromatography, Sigma L3024), was administered 2 hrs before the experiment, at a dose of 1.5 × 10⁶ EU/kg body weight, in 0.5 ml of sterile saline, as previously described (3). In some experiments, sCD40L (Peprotech Inc., Rocky Hill, NJ, 10ng/mouse) was administered intravenously at the time of LPS injection.

Cremaster muscle preparation

Previous studies from our laboratory have demonstrated that DSS colitis is associated with enhanced light/dye-induced thrombus formation in the microvasculature of mouse cremaster muscle (3,4,12). This reproducible response to DSS was manifested largely in arterioles but not in venules. We (3) and others (6, 13) have shown that light/dye-induced thrombosis in cremaster muscle is also enhanced by LPS administration, but the response is largely manifested in postcapillary venules. The cremaster muscle preparation was used in this study to evaluate the role of CD40/CD40L in the enhanced microvascular thrombosis that accompanies either DSS colitis or LPS. The cremaster muscle was prepared using mice anesthetized with sodium pentobarbital (50mg/kg) administered by intraperitoneal injection. Following a scrotal incision, the muscle was carefully separated from the testicle and epididymis, positioned on a temperature-controlled pedestal and suffused continuously (1 ml/min) with bicarbonate-buffered saline (pH 7.35-7.45). Muscle and core body temperatures were maintained at 37 ± 0.5°C.

Brain preparation

Endotoxin-enhanced microvascular thrombosis was also studied in the brain of WT, CD40^-/- and CD40L^-/- mice. For these experiments, the cerebral microvasculature was observed via a cranial window, as previously described (14). The mice were anesthetized as described above, tracheostomized and artificially ventilated with room air. Following the craniectomy, the dura was carefully reflected. The brain surface was bathed in artificial cerebrospinal fluid (in mEq/L): 147.8 Na⁺, 3.0 K⁺, 2.3 Mg²⁺, 2.3 Ca²⁺, 135.2 Cl^-, 19.61 HCO₃^-, 1.67 lactate.

Light/dye thrombosis model

The in situ tissue preparations (brain and cremaster muscle) were transferred onto the stage of an upright intravital fluorescence microscope equipped with a 175-W xenon light source and allowed to equilibrate for 30 min. A 20x objective was used to visualize the microcirculation, while individual vessels subjected to the thrombosis procedure were observed with a 40x water immersion objective. A medium-resolution video camera (Dage MTI CCD-72) projected the images onto a monitor (Sony TRINITRON PVM-2030, Japan) and recorded on a DVD player (Panasonic DMR-E50, NJ). A video time/date generator projected the time, date, and stopwatch function on the monitor (diagonal 50.6 cm, PVM-2030, Sony Trinitron; Japan).

After the equilibration period, 10mg/kg of 5% FITC dextran (150,000 mol wt, Sigma Chemicals, St Louis) was injected intravenously via a jugular vein and allowed to circulate for approximately 10 min before photoactivation of the first vessel (6,13). In the brain, branches of postcapillary venules and arterioles that drain blood from the anterior, superior, and middle cerebral arteries were chosen with diameters of between 30 μm and 50 μm. In the cremaster, second- or third-order unbranched venules and arterioles with diameters ranging between 30 and 50μm were selected for study. Photoactivation was initiated by exposing approximately 100 μm of vessel length to epi-illumination, with a 175-W xenon lamp (Lamda LS, Sutter) and a fluorescein filter cube (exciter-barrier filter and reflector combination, blue excitation 450-490, Chroma). The excitation power density was measured daily (IL 1700 Radiometer, SED-033 detector, International Light) and maintained within 1% of the target of 1.4 W/cm², as described previously (3, 6). Epi-illumination was applied continuously. In the cremaster muscle preparation, thrombus formation was quantified by determining (1) time to onset of visible aggregates (onset time) and (2) the time required for complete flow cessation for ≥60s (cessation time). Since the image resolution of microvessels in brain tissue did not allow for an accurate determination of onset time, thrombus formation in this tissue was monitored using only cessation time. Typically, 2–4 thrombi were induced in each mouse, and the results of each vessel type (venules, arterioles) were averaged.

Immunohistochemistry

Mice (WT and CD40-/-, saline and LPS treated) were anesthetized with overdose of ketamine (100 mg/kg) and intracardially perfused with 10 ml phosphate buffered saline (PBS) followed by 20 ml of 4% paraformaldehyde (PFA) in PBS. Following perfusion, brains and cremaster muscle were dissected and post-fixed in 4% PFA in PBS for 4 h. The fixed tissue samples were transferred to 10–20–30% sucrose (w/v) in PBS until equilibrated (24 h, 4°C). Tissue samples were then frozen in Tissue-Tek OCT mounting medium, and 10 μm-thick sections were cut with a cryostat. Sections were subsequently mounted onto poly-l-lysine-coated slides and allowed to air dry and stored in −80 °C until used. Brain sections were collected between bregma −0.8 and 1.70 mm. Non-specific binding sites were blocked with 10% normal serum and Avidin-blocking solution (Vector Laboratories). Sections were incubated with purified anti-mouse CD40 antibody (BioLegend, 1:300 diluted in 10% normal serum and biotin-blocking solution in PBS) at 4 °C overnight. Sections were then incubated with biotinylated goat anti-rat IgG secondary antibody (1:200 in PBS; Vector Laboratories) and signals visualized by streptavidin conjugated Alexa Fluor 488. Sections were mounted with hard set mounting medium (Vector lab H-1400). For control staining, primary antibody was replaced by rat IgG (I-4000, Vector Laboratories). All controls gave negative results with no detectable labeling.

Statistical analysis

All values are reported as means ± SEM. ANOVA with the Fisher post hoc test was used to compare groups, with statistical significance set at P < 0.05.

Results

A total of 152 mice were used in this study, with 42 and 110 mice used for the cerebral and cremaster muscle experiments, respectively. The n values for each experimental group (for each vessel type) are provided in the figure legends. The average body weight of all mice in this study was 26.31 ± 0.30 g. WT mice placed on the DSS regimen exhibited a significant reduction in body weight (22.6 ± 1.0 g) and this reduction was also seen in the mutant mice. The average diameters of the vessels studied were: 37.1 ± 0.73 μm for arterioles and 34.26 ± 0.74 μm for venules in brain, and 40.74 ± 0.52 μm for arterioles and 36.05 ± 0.42 μm for venules in cremaster muscle.

The responses (time to flow cessation) of the cerebral microvasculature in wild type control (saline) group exhibiting a flow cessation time of 12 ± 0.7 min in venules, are summarized in Figure 1. As previously reported (3, 6), thrombus development occurs more rapidly in venules than arterioles following light/dye injury, with the control (saline) group exhibiting a flow cessation time of 12 ± 0.7 min in venules, compared to 31.4 ± 1.4 min in arterioles. LPS administration significantly accelerated thrombus formation in both venules and arterioles, with the former vessels exhibiting a more dramatic reduction in cessation time (52%) than the latter (33%).

Figure 1. Endotoxin enhanced light/dye-induced thrombosis in cerebral venules and arterioles of wild type mice, CD40 deficient (CD40^-/-) miceand CD40L deficient (CD40L^-/-) mice.

Values represent time to flow cessation following light/dye injury in (A) arterioles and (B) venules. Soluble CD40L (10ng/mouse) was administered intravenously at the time of LPS injection 10ng/mouse) to some mice. The numbers of mice per experimental group: venules-saline = 6, venules-LPS = 6, arterioles-saline = 6 and arterioles-LPS = 6. *p<0.05 vs. corresponding saline group. #p<0.05 vs. corresponding sCD40L-saline group.

Figure 1 also compares the light/dye-induced thrombosis responses in cerebral venules and arterioles of CD40 deficient (CD40-/-) mice treated with either saline or LPS. Cerebral venules in saline-treated CD40-/- mice exhibited a significantly delayed thrombosis response (29.0 ± 3.7 min), when compared to the control (saline) response noted in WT mice (12 ± 0.7 min). LPS treatment greatly accelerated thrombus formation in venules of CD40-/- mice, when compared to their saline-treated counterparts; although the LPS value in CD40-/- mice did not differ from the LPS value in WT mice. A more modest prolongation of thrombus formation was noted in cerebral arterioles of CD40-/- mice treated with saline (compared to WT mice: 40.43 ± 2.78 min vs. 31.43 ± 1.38 min respectively). However, LPS-enhanced thrombosis was not observed in arterioles of CD40-/- mice.

In CD40L-/- mice, the time to flow cessation in cerebral venules was only marginally increased following saline treatment, when compared to their WT counterparts (Figure 1). However, LPS did not enhance thrombus formation in these vessels (a similar response was observed in cremaster muscle, as depicted in Figure 2). A dramatic delay (59.0 ± 1.7 min) of thrombus formation was noted in arterioles of saline-treated mice of CD40L-/- mice, when compared to the response observed in WT mice (31.4 ± 1.4 min). Although LPS significantly enhanced thrombosis in arterioles of CD40L-/- mice, this response (46.2 ± 1.7 min) was significantly blunted when compared to WT mice receiving LPS (19.6 min ± 1.0 min).

Figure 2. Effects of endotoxin (LPS) on light/dye-induced thrombosis responses in cremaster muscle arterioles and venules of wild type (WT), CD40^-/- and CD40L^-/- mice.

Values are presented for time of onset of thrombus formation (onset) and time to flow cessation in (A) arterioles and (B) venules. The numbers of mice per experimental group: arterioles, WT-control = 15, WT-LPS = 7, CD40L^-/- = 7, CD40L^-/--LPS = 5, CD40^-/- = 9, CD40^-/--LPS = 8; venules, WT-control = 18, WT-LPS = 10, CD40L^-/- = 10, CD40L^-/--LPS = 7, CD40^-/-= 10, CD40^-/--LPS = 8. *p<0.05 vs. corresponding WT-control group, ^†p<0.05 vs. corresponding WT-LPS group, #p<0.05 vs. corresponding CD40^-/--LPS group.

Figure 1 also summarizes the responses of cerebral venules and arterioles to LPS administration in CD40L-/- mice receiving an intravenous dose of soluble CD40L. In both vessel types, sCD40L administration reversed the beneficial effects afforded by CD40L deficiency. No effect was observed in saline treated CD40L-/- mice receiving soluble CD40L. The reduced times to flow cessation observed in venules and arterioles of CD40L-/- mice treated with sCD40L were comparable to the values noted in LPS treated WT mice.

We also determined whether the effects of CD40- or CD40L-deficiency observed in the brain microvasculature are also manifested in a peripheral vascular bed, such as cremaster muscle. The thrombosis responses elicited by light/dye injury alone (in the absence of an external inflammatory stimulus) in cremaster microvessels of WT, CD40-/- and CD40L-/- mice are summarized in Figure 2. The time of onset of thrombus formation in both arterioles and venules was not affected by genetic deficiency of either CD40 or CD40L. As previously reported for WT mice (3, 5), LPS causes a significant increase in thrombus development in venules in comparison to arterioles, in which little or no significant effect is seen. Whilst an LPS enhanced thrombus development in venules, no change in blood flow cessation time was noted in arterioles. A significant prolongation of vessel occlusion time was noted in LPS-treated arterioles of CD40L-/- mice, while CD40-/- mice responded in a similar fashion as WT mice to LPS. However, when compared to their saline controls, LPS did not alter blood flow cessation time in both CD40-/- and CD40L-/- mice. In venules, CD40 (but not CD40L) deficiency significantly prolonged the time to flow cessation, when compared to LPS- treated WT mice.

Immunohistochemistry was performed to identify CD40 expression in cerebral and cremaster muscle of both saline and LPS treated WT mice and CD40-/- mice (control). Figure 3 shows the increase in CD40 expression in mice treated with LPS in both brain (figure 3A & B) and cremaster muscle (figure 3C & D). As expected, little or no staining was observed in CD40-/- mice or with the IgG isotype control

Figure 3. CD40 expression in cremaster muscle and brain of WT mice.

LPS (dose of 1.5 × 10⁶ EU/kg body weight, in 0.5 ml of sterile saline,) or saline was injected i.p for 2 hours. Brain and cremaster muscle were then removed. CD40 staining was performed on: (A) cerebral sections and (B) cremaster muscle from saline treated WT mice and (C) cerebral sections and (D) cremaster muscle from LPS treated WT mice. Bar = 50 μm.

Figure 4 further demonstrates the differences between vascular beds, such as the brain and cremaster muscle. This figure summarizes the responses of cremasteric venules and arterioles to LPS administration in CD40L-/- mice receiving an intravenous dose of soluble CD40L. Unlike in the brain, sCD40L administration reversed the beneficial effects afforded by CD40L deficiency in arterioles only. This reduced time to flow cessation quantified in arterioles of CD40L-/- mice treated with sCD40L were similar to those observed in LPS treated WT mice.

Figure 4. Effects of soluble CD40L in cremaster muscle arterioles and venules of CD40L^-/- mice.

Values are presented for time of onset of thrombus formation (onset) and time to flow cessation in (A) arterioles and (B) venules. The numbers of mice per experimental group: arterioles, WT-control = 18, WT-LPS = 10, CD40L^-/- = 7, CD40L^-/--LPS = 5, CD40^-/- = 9, CD40^-/--LPS = 8; venules, WT-control = 22, WT-LPS = 13, CD40L^-/- = 10, CD40L^-/--LPS = 10, CD40^-/-= 10, CD40L-/--LPS + sCD40L = 5, CD40^-/--LPS = 8. *p<0.05 vs. corresponding WT-control group, ^†p<0.05 vs. corresponding WT-LPS group, #p<0.05 vs. corresponding CD40^-/--LPS group.

The effects of CD40 and CD40L deficiency on light/dye-induced thrombus formation in cremaster arterioles and venules of DSS colitic mice are summarized in Figure 5. As previously reported (3), DSS colitis enhances both the initiation (time of onset) and propagation/stabilization (time to flow cessation) phases of thrombus formation in cremaster muscle arterioles, but not in venules. Blood flow cessation time was significantly longer in CD40L-/- DSS colitic mice versus either WT-DSS or CD40-/- DSS mice, but considerably less than CD40L-/- control mice. CD40 deficiency appeared to further accelerate the onset (p<0.05) of thrombus development without significantly (p = 0.34) altering the time to flow cessation.

Figure 5. Effects of dextran sodium sulphate (DSS) colitis on light/dye-induced thrombosis responses in cremaster muscle arterioles and venules of wild type (WT), CD40^-/- and CD40L^-/- mice.

Values are presented for time of onset of thrombus formation (onset) and time to flow cessation in (A) venules and (B) arterioles. The numbers of mice per experimental group: arterioles, WT-control = 15, WT-DSS = 6, CD40^-/- = 9, CD40^-/--DSS = 6; CD40L^-/- = 7, CD40L^-/--DSS = 6; venules, WT-control = 18, WT-DSS = 7, CD40^-/-= 10,CD40^-/--DSS = 6, CD40L^-/- = 10, CD40L^-/--DSS = 7. *p<0.05 vs. corresponding WT-control group, ^†p<0.05 vs. corresponding WT-DSS group, #p<0.05 vs. corresponding CD40L^-/--DSS group, $p<0.05 vs. corresponding CD40^-/--DSS group.

Discussion

The recognition that many pathological conditions are associated with both inflammation and coagulation has led to the concept that these are two interdependent processes that can propagate and intensify each other (1). However, the factors that link inflammation to coagulation and thrombosis remain poorly understood. One signalling pathway that has been implicated in both inflammation and coagulation is CD40/CD40L. While this pathway has been linked to the thrombosis associated with vessel injury in both large arteries and arterioles (7-9), no previous effort has been made to determine whether it contributes to the accelerated thrombus formation that is observed in the microvasculature during either acute or chronic inflammation. Using two distinct models of inflammation, we provide evidence that implicates CD40 and/or CD40L in inflammation-enhanced microvascular thrombosis in the cerebral and muscle microvasculatures. Our findings also reveal that the relative contributions of CD40 and its ligand to inflammation-enhanced thrombosis differ between vascular beds (brain vs. muscle) and vessel types (arterioles vs. venules).

The CD40/CD40L dyad has been implicated in the regulation of adhesion molecule expression, production and release of reactive oxygen species, cytokines and other inflammatory mediators (such as TNFα IL-1α and β and IL-6), and in the activation of coagulation factors (10). CD40 is expressed on vascular endothelial cells and its expression is increased by acute and chronic inflammatory stimuli, such as LPS (15) and DSS-induced colitis (11). In the present study, we demonstrated that LPS was indeed able to induce CD40 expression in both the brain and cremaster muscle of WT mice. As expected, the CD40-/- mice did not exhibit detectable levels of CD40 expression. CD40L, a membrane glycoprotein belonging to the tumour necrosis factor family, is expressed by activated T-cells and platelets. CD40L is shed from the cell surface (predominantly from platelets) and its concentration in plasma has been used as a prognostic marker for inflammation and thrombotic risk in human disease (10, 16, 17). The soluble circulating form of CD40L retains biological activity and is able to engage and activate its receptor on endothelial cells and T-cells (18). The binding of soluble or cell membrane-bound CD40L to endothelial cells induces a prothrombogenic phenotype, which results in part from an upregulation of tissue factor (19). Studies of large and microscopic arterial vessels suggest that CD40L may be essential for thrombus stabilization (7-9).

Our evaluation of LPS-enhanced thrombosis in the cerebral microcirculation reveals a differential role for CD40 and CD40L in arterioles and venules. In brain arterioles, genetic deficiency of CD40 completely, while CD40L deficiency only partially, prevents the accelerated thrombosis induced by LPS administration. In venules, on the other hand, the LPS-induced enhancement of thrombus formation was completely prevented in CD40L-/- mice, while CD40 deficiency afforded no significant protection. CD40 expression was significantly increased in WT mice treated with LPS in both the brain and the cremaster muscle vascular beds. Although further studies are needed to elucidate the role of the CD40/CD40L dyad in LPS-enhanced thrombosis, the possibility remains that CD40L participates in the thrombosis response largely due to its ability to interact with other adhesion glycoproteins that are expressed on platelets and/or leukocytes, such as CD11b, GPIIbIIIa, and α₅β₁ (19-21).

A novel and potentially important observation in the present study was the ability of exogenous soluble CD40L to reverse the protection against LPS-enhanced thrombus formation in cerebral arterioles and venules that was afforded by CD40L deficiency. This observation suggests that the protection observed in CD40L-/- mice may reflect the absence of the soluble, rather than the membrane-bound, form of the ligand. Furthermore, the accelerated LPS-enhanced thrombosis observed following intravascular administration of sCD40L, particularly in the brain, suggests that the elevated sCD40L levels detected in septic patients may contribute to the increased thrombotic risk in this disease (22). Other studies by our group have shown a tissue-specific, time-dependent upregulation of endothelial CD40 in mice challenged with bacterial LPS, which likely reflects organ-to-organ differences in the responsiveness of endothelial cells and perivascular auxiliary cells, such as macrophages and mast cells, to LPS-induced activation (15).

Endothelial cells within the brain have a unique expression pattern of cell surface receptors, transporters, and intracellular enzymes, which regulate solute exchange between the blood and the brain (23,24). Other studies demonstrate vascular bed-specific endothelial cell responses, such as those elicited by LPS, which can upregulate certain genes, e.g. ptx-3 (a member of the pentraxin gene family) specifically within heart and skeletal muscle vascular beds (25).

Previous reports have described the differential effects of LPS administration vs. DSS-induced colitis on thrombus formation in venules and arterioles of the cremaster muscle (3, 6). Like these studies, we show here that LPS enhances light/dye-induced thrombosis only in venules, while DSS colitis preferentially targets the arterioles. These observations not only suggest that endogenous LPS is an unlikely mediator of the enhanced thrombus formation associated with experimental colitis, but also highlight the ability of different inflammatory stimuli to preferentially promote thrombus formation in specific segments of the microvasculature. A comparison of the LPS-induced responses between cremaster muscle and brain also reveals the tissue-specificity of thrombus development to a given inflammatory stimulus. In the cerebral microvasculature, a significant enhancement of thrombus development is observed in both arterioles and venules, while only venules respond to LPS in cremaster muscle. Similarly, the LPS-mediated responses of cerebral and muscle venules to CD40 deficiency were remarkably different, with the former showing an unaltered response compared to their WT counterparts while the LPS-enhanced thrombosis was completely reversed in the latter vessels.

CD40/CD40L signalling has been implicated in the pathogenesis of human and experimental colitis (10, 18). Mutant mice that are deficient in either CD40 or CD40L have been shown to be equally effective in blunting the gut inflammatory and injury responses to DSS, and in reducing the recruitment of adherent platelets and leukocytes in inflamed colonic venules (11), supporting a role for ligand-receptor mediated signalling. In the present study of DSS-enhanced light/dye-induced thrombosis in cremaster muscle arterioles, we noted significant protection against thrombus formation by CD40L deficiency, while CD40-/- mice behaved like their WT counterparts. These observations suggest that CD40L exerts its effects in colitis by engaging with a receptor or protein other than CD40, possibly the platelet receptor GPIIb/IIIa (20). It appears unlikely that CD40L deficiency affords protection by blunting DSS-induced gut inflammation (and reducing the circulating signal that enhances thrombosis in distal tissue) because we have already shown similar attenuation of gut inflammation in CD40-/- and CD40L-/- mice (11). A possible explanation for the different thrombotic responses in CD40-/- and CD40L-/- mice is that the shedding of CD40L from the large number of adherent (and presumably activated) platelets in inflamed colonic venules may result in an increased circulating level of sCD40L, as previously reported for human colitis (16, 26), which could subsequently contribute to the enhanced thrombosis response observed in cremaster muscle.

In conclusion, the results of this study implicate CD40 and its ligand CD40L in the enhanced light/dye-induced microvascular thrombosis associated with an acute (LPS) and chronic model (DSS colitis) of inflammation. The responses and contributions of CD40 and CD40L to the thrombosis responses elicited by these models exhibit some specificity relative to vascular bed (brain vs. cremaster muscle) and vascular segment (arteriole vs. venule). The protection against thrombosis that afforded by genetic deficiency of either CD40 or its ligand support the development of therapeutic strategies that target either or both components of CD40/CD40L signalling to reduce thrombotic risk associated with acute and chronic inflammatory diseases.