New Insights into the Molecular Basis of the Antiphospholipid Syndrome

Abstract

Antiphospholipid syndrome is an autoimmune disease characterized by the presence of circulating antiphospholipid antibodies (aPL) that promote thrombosis, pregnancy complications and cardiovascular diseases. Alterations in the function of vascular cells induced by aPL underlie these outcomes. This review will discuss recent findings that indicate a novel mechanism by which aPL antagonize endothelial cell production of nitric oxide and thereby promote thrombosis.

Introduction

The antiphospholipid syndrome (APS) is an autoimmune disease characterized by the production of antiphospholipid antibodies (aPL) that promote vascular thrombosis and pregnancy loss^{1, 2}. In addition to its occurrence in individuals without an underlying disorder, APS afflicts a significant number of patients with systemic lupus erythematosus (SLE), with as many as 34% of lupus patients having circulating aPL¹. Along with arterial and venous thrombosis and pregnancy complications, patients with APS have an increased risk of coronary artery disease, myocardial infarction, and stroke³. A link between APS and premature atherosclerosis has also been reported^{4, 5}.

Human studies as well as work in cell culture and in animal models indicate that actions of aPL on endothelial cells likely play a major role in the vascular disease phenotypes in APS^{6, 7}. There is evidence of endothelial cell activation in APS patients, with plasma levels of soluble adhesion molecules such as vascular cell adhesion molecule-1 (VCAM-1), intercellular adhesion molecule-1 (ICAM-1) and von Villebrand factor (vWF) increased in APS patients compared to healthy controls^8–10. In addition, elevations in circulating endothelial microparticles and in circulating mature endothelial cells, which are markers of endothelial activation and damage, have been observed in APS patients^{8, 11}. Several studies also indicate that flow-mediated vasodilation is blunted in APS patients compared to healthy subjects^{8, 10, 12}. Consistent with these findings in humans, the exposure of cultured endothelial cells to aPL isolated from APS patients causes VCAM-1, ICAM-1 and E-selectin upregulation, and it also increases the expression of tissue factor (TF). Similarly, in mouse models the administration of aPL causes increased expression of adhesion molecules and it enhances endothelial cell-leukocyte interaction^{13, 14}. Furthermore, mirroring the human condition, the injection of aPL in rodent models leads to enhanced thrombus formation^{15, 16}

Over the past two decades our knowledge of the pathogenetic mechanisms underlying APS have been expanded through studies of the molecular pathways by which aPL alter the function of endothelium and platelets^{7, 17}. This review will focus on recent findings that indicate a novel mechanism by which aPL antagonize endothelial cell production of nitric oxide (NO) and thereby promote thrombosis.

Endothelial NO Synthase Antagonism by aPL

One of the key signaling molecules that has a beneficial impact on vascular health by preventing thrombosis and endothelial cell-leukocyte interaction is NO^{18, 19}. The primary source of NO in the vascular wall under normal conditions is the endothelial isoform of NO synthase (eNOS). The NO generated by eNOS downregulates adhesion molecule expression, and it also inhibits platelet aggregation by increasing cGMP production in platelets^{18, 20}. Thus, the activation of eNOS and subsequent production of NO modulate a number of the vascular processes that are known to be adversely affected by aPL. Studies in both mouse models and humans have suggested that there is a potential link between aPL and changes in bioavailable NO. In mice the administration of aPL reduces plasma concentrations of NO metabolites and it also reduces acetylcholine (Ach)-induced relaxation in isolated aortic rings, which is an endothelium-dependent, NO-dependent process^{21, 22}. In humans, plasma aPL levels are inversely correlated with urinary NO metabolite excretion, and APS patients have lower levels of plasma nitrites compared to control subjects^{23, 24}. Thus, there are data both in mouse models and in humans that support a potential role for impaired NO production in the pathogenesis of APS.

To directly test this possibility and determine the underlying processes, Ramesh et al. recently determined if aPL alter eNOS activation in cultured endothelial cells²⁵. Human or bovine aortic endothelial cells were pretreated with polyclonal aPL or normal human IgG (NHIgG) isolated from APS patients or healthy individuals, respectively, and eNOS activation by vascular endothelial growth factor (VEGF) was evaluated. In the presence of NHIgG, VEGF treatment led to a predictable increase in eNOS activity. In contrast, aPL caused complete attenuation of eNOS activation by VEGF. The activation of the enzyme by other agonists was also blunted by aPL. To test whether these mechanisms are operative in vivo, the investigators evaluated carotid vascular conductance responses to Ach in mice, which indicate increases in blood flow invoked by endothelium-derived NO²⁶. Ach-mediated increases in carotid artery vascular conductance were tested at baseline and following the IV administration of either NHIgG or aPL. Whereas NHIgG did not alter the response to Ach, the administration of aPL caused marked attenuation of the vasodilatory response. These findings indicate that aPL have a dramatic inhibitory effect on eNOS both in cell culture and in vivo in mice.

Since NO plays an important role in preventing leukocyte adhesion to endothelium and it is antithrombotic²⁰, Ramesh et al. further determined the implications of eNOS antagonism of aPL in studies of leukocyte-endothelial cell adhesion and thrombus formation using intravital microscopy of the mesenteric microcirculation. Wild-type (eNOS^+/+) or eNOS deficient (eNOS^−/−) mice were treated with NHIgG or aPL, and 24 hours later endothelial cell-leukocyte adhesion and thrombosis were studied. In eNOS^+/+ mice aPL caused a marked increase in leukocyte adhesion and thrombus formation. In contrast, in eNOS^−/− mice aPL did not affect either leukocyte adhesion or thrombosis. These combined results indicate that eNOS antagonism underlies the increase in both leukocyte-endothelial cell adhesion and thrombus formation caused by aPL.

Ramesh et al. further determined how aPL blunt the stimulation of eNOS, which is caused by most agonists via the activating phosphorylation of Ser1177 of the enzyme (Ser1177 in human eNOS, Ser1179 in bovine eNOS)^{27, 28}. Whereas in the presence of control NHIgG VEGF caused increased phosphorylation of eNOS at Ser1177, aPL completely suppressed VEGF-induced Ser1177 phosphorylation. It was additionally determined that this process does not involve alterations in VEGF-stimulated Akt kinase activation, which is responsible for Ser1177 phosphorylation. Alternatively, the impaired phosphorylation of eNOS caused by aPL was mediated by the protein phosphatase 2A (PP2A), which is the principal phosphatase that dephosphorylates eNOS at Ser1177^{29, 30} This distal event in the actions of aPL on eNOS is depicted in Figure 1(c).

Figure 1. aPL promote leukocyte adhesion and thrombosis by antagonizing eNOS viaβ2GPI, apoER2 and the phosphatase PP2A.

(a) aPL binding to domain I of β2GPI induces β2GPI dimerization and (b) interaction between domain V of β2GPI and the first LDL binding domain of apoER2. Through yet-to-be-determined mechanism(s), the interaction of β2GPI with apoER2 causes (c) increased activation of PP2A. This promotes the dephosphorylation of Ser1177 of eNOS yielding decreased enzyme activity and a decline in bioavailable NO, which results in increased leukocyte adhesion and increased thrombosis. Reprinted with permission ²⁵.

Role of β2-Glycoprotein I

Although initially thought to directly recognize anionic phospholipids on the surface of target cells, it has become apparent that pathologically relevant aPL are directed against phospholipid-binding proteins, particularly β2-glycoprotein I (β2GPI) ^{31, 32}. β2GPI is a plasma protein composed of five distinctive domains (domains I-V), and it has been shown that it binds to phospholipids on the surface of platelets through domain V³². In vitro studies have shown that antibody recognition of phospholipid-bound β2GPI causes its dimerization, which further increases its affinity for negatively-charged phospholipids on the cell surface³³. Anti-β2GPI-specific monoclonal antibodies increase adhesion molecule expression and the synthesis of cytokines, endothelin-1, and TF in cultured endothelial cells, and in isolated platelets they enhance the production of thromboxane B₂, adhesion to collagen and aggregation ^{14, 34}. In APS patients there is a strong relationship between the levels of circulating anti-β2GPI antibodies specific for domain I and the incidence of thrombosis^{35, 36}.

Ramesh and colleagues evaluated the involvement of β2GPI in aPL antagonism of eNOS in cultured endothelial cells²⁵. When cells were deprived of serum, which is the source of β2GPI in culture, exposure to aPL did not cause eNOS inhibition, indicating that β2GPI is required for aPL action. The role of β2GPI was further confirmed by testing the effect of anti-β2GPI monoclonal antibodies. Monoclonal anti-β2GPI antibodies that recognize domain I inhibited VEGF-induced eNOS activation, mimicking the effect of the polyclonal aPL that is isolated from APS patients; in contrast, a monoclonal antibody that recognizes domain II of β2GPI did not. These findings indicate that β2GPI is required for aPL actions on eNOS and that aPL recognition of domain I of β2GPI mediates the process. Further experiments revealed that it is not simply the recognition of domain I of β2GPI by aPL that initiates the process; the resulting dimerization of plasma membrane-associated β2GPI is also required (Figure 1(a)).

Role of apolipoprotein E receptor 2

Previous in vitro studies found that dimerized β2GPI binds to multiple members of the LDL receptor family in purified form^{34, 37}. One such receptor is apolipoprotein E receptor 2 (apoER2, also known as LRP8), which plays a critical role in brain development as a signal tranducer for the glycoprotein Reelin³⁸. Although apoER2 is predominantly expressed in brain, a splice variant of apoER2 designated apoER2' has been identified in platelets and megakaryocytic cell lines³⁴. Recently it has been demonstrated that aPL induce β2GPI binding to apoER2' in isolated platelets, and that the binding causes platelet activation ex vivo³⁴.

Recognizing that apoER2 is an abundant plasma membrane receptor in endothelial cells and that it interacts with dimerized β2GPI, Ramesh and coworkers tested the requirement for apoER2 in aPL-induced eNOS antagonism²⁵. In cultured endothelial cells, the knockdown of apoER2 expression by small interference RNA (RNAi) had no effect on the promotion of monocyte-endothelial cell adhesion prompted by lipopolysaccharide (LPS) treatment of the endothelium. However, it fully prevented the enhancement of adhesion caused by anti-β2GPI antibody. eNOS antagonism by aPL was also fully prevented by apoER2 knockdown. Studies with a soluble peptide based on the sequence of the first LDL-binding domain of apoER2’ (BD1) that binds to β2GPI³⁹ further demonstrated that the eNOS antagonism caused by antibody recognition of β2GPI requires interaction between β2GPI and BD1 of the receptor. As such, apoER2 is the linchpin that couples antibody recognition of β2GPI and its dimerization on the endothelial cell surface to the antagonism of intracellular eNOS and the changes in endothelial cell behavior that ensue (Figure 1(b)).

Based on these observations in cell culture, Ramesh et al. evaluated the role of apoER2 in aPL actions in vivo in a series of experiments in apoER2^+/+ and apoER2^−/− mice (Figure 2). The requirement for the receptor in aPL-mediated antagonism of eNOS in vivo was tested in studies of carotid vascular conductance (Figure 2A, B). aPL caused marked attenuation of the vasodilatory response to Ach in apoER2^+/+ mice, shifting the dose-response curve downward and to the right. In contrast, Ach-induced vasodilation was identical before and after aPL treatment in apoER2^−/− mice. The involvement of apoER2 in aPL-induced increases in leukocyte adhesion to endothelium and in thrombus formation was also evaluated (Figure 2C-F; representative still images in 2C and 2E, quantitative data in 2D and 2F). ApoER2^+/+ or apoER2^−/− mice were treated with NHIgG or aPL, and 24 hours later endothelial cell-leukocyte adhesion or thrombus formation was evaluated using intravital microscopy. In apoER2^+/+ mice, aPL increased leukocyte adhesion to endothelium resulting in a decrease in leukocyte velocity (Figure 2C, D). In contrast, aPL had no effect on leukocyte-endothelial cell adhesion in apoER2^−/− mice. Similarly, in apoER2^+/+ mice, compared to NHIgG, aPL enhanced thrombus formation and thereby decreased the time to total vessel occlusion (Figure 2E, F). In contrast, aPL did not alter thrombus formation in apoER2^−/− mice. Romay-Panabad et al. recently confirmed the role of apoER2 in aPL-induced thrombosis⁴⁰. They found that human polyclonal aPL or murine anti-β2GPI monoclonal antibody enhances TF production in the carotid artery and increases thrombus formation in wild-type mice, and that these pathologic effects of antibody are reduced in apoER2^−/− mice. Furthermore, they showed that the blocking peptide that interferes with β2GPI-apoER2 interaction attenuates these pathologic events invoked by aPL or anti-β2GPI antibody in wild-type mice. These cumulative observations indicate that aPL recognition of β2GPI causes eNOS antagonism, that this process is mediated by apoER2, and that it underlies the vascular disease phenotypes of APS.

Figure 2. aPL-induced eNOS antagonism, endothelial cell-leukocyte adhesion and thrombosis in mice are mediated by apoER2.

(A, B) ApoER2^+/+ and apoER2^−/− mice were instrumented and changes in carotid vascular conductance in response to acetylcholine were compared before and after treatment with aPL. Dose-responses to acetylcholine were performed sequentially at baseline (●), 60 min after aPL treatment (○), and 10 min after L-NAME administration (▼). n=6/group. *p<0.05 vs. baseline dose-response. (C-F) ApoER2^+/+ and apoER2^−/− mice were injected IP with NHIgG or aPL, and 24 hours later prepared for intravital microscopy. Leukocytes or platelets were fluorescence-labeled by injecting Rhodamine 6G or anti-mouse GPIbβ antibody, respectively. The mesentery was exposed for the observation and recording of images of leukocyte rolling or thrombus formation. Representative still images are shown (C, E). (In C, leukocytes are indicated by arrows). Quantitative data are shown as velocity of leukocyte rolling (D) and time required for complete vessel occlusion (F) in apoER2^+/+ and apoER2^−/− mice. In D and F, values are mean±SEM, n=5–7/group, *p<0.05 vs. NHIgG, †p<0.05 vs. apoER2^+/+. Reprinted with permission²⁵ .

Conclusions

Patients with APS suffer from arterial and venous thrombosis, and they are at increased risk of myocardial infarction and stroke. Recent experiments in cell culture and in mice have provided new knowledge of the underlying events, revealing that aPL inhibit eNOS via recognition of the cell surface protein β2GPI and its interaction with the plasma membrane receptor apoER2. However, numerous aspects of the molecular mechanisms of aPL action remain unknown. How does apoER2 activate the phosphatase PP2A? Does a known adaptor molecule of apoER2, such as Disabled-1, play a role in the signal transduction^{38, 41}? Do similar processes occur in platelets, which express the splice variant of apoER2, apoER2’, and eNOS, which attenuates platelet activation in an autocrine manner ⁴²? Perhaps most importantly, how does aPL recognition of β2GPI and resulting signaling by apoER2 occur in a manner that would explain the episodic nature of the thrombotic diathesis that characterizes APS? Because APS patients most often do not display clinical symptoms even in the persistent presence of circulating aPL, it has been hypothesized that additional stimuli are required for the occurrence of thrombosis episodes (designated the “two-hit hypothesis”)⁴³. If impaired NO production and subsequent endothelial dysfunction induced by aPL represent "the first hit", what is the “second hit” that triggers the thrombotic events in APS? The “second hit” may involve a proinflammatory stimulus or infection because rats administered aPL have spontaneous thromboses if they also receive LPS⁴⁴. Furthermore, how are the “first hit” and the “second hit” mechanistically coupled? As the molecular mechanisms responsible for the “first hit” are further revealed, the nature of the other required processes that ultimately lead to the thrombotic events will become more approachable.

Currently patients with APS are treated chronically with anticoagulation medications such as heparin or warfarin. Despite long-term use of the anticoagulants, recurrent thrombosis can occur in APS patients. Furthermore, use of anticoagulants is fraught with complications including bleeding episodes and osteopenia^45–47. With the recent new knowledge that has been gained regarding the processes that underlie aPL actions on endothelium, we anticipate that novel interventions can be developed that directly target the pathogenetic mechanisms, thereby affording greater efficacy and fewer complications in the management of this potentially life-threatening disorder.