To investigate the ability of statins, the inhibitors of the hydroxymethylglutaryl–coenzyme A reductase enzyme, to affect endothelial cell activation induced by anti–β2-glycoprotein I (anti-β2GPI) antibodies in vitro.
To investigate the ability of statins, the inhibitors of the hydroxymethylglutaryl–coenzyme A reductase enzyme, to affect endothelial cell activation induced by anti–β2-glycoprotein I (anti-β2GPI) antibodies in vitro.
Human umbilical vein endothelial cell (HUVEC) activation was evaluated as U937 monocyte adhesion, E-selectin, and intercellular adhesion molecule 1 (ICAM-1) expression by cell enzyme-linked immunosorbent assay and as interleukin-6 (IL-6) messenger RNA (mRNA) expression by RNA protection assay. E-selectin–specific nuclear factor κB (NF-κB) DNA-binding activity was evaluated by the gel-shift assay. HUVECs were activated by polyclonal affinity-purified IgG, human monoclonal IgM anti-β2GPI antibodies, human recombinant IL-1β, tumor necrosis factor α, or lipopolysaccharide (LPS).
Fluvastatin reduced, in a concentration-dependent manner (1–10 μM), the adhesion of U937 to HUVECs and the expression of E-selectin and ICAM-1 induced by anti-β2GPI antibodies as well as by cytokines or LPS. Another lipophilic statin, simvastatin, display similar effects but to a lesser extent than fluvastatin. The inhibition of E-selectin expression exerted by fluvastatin was related to the impairment of NF-κB binding to DNA. Moreover, the drug attenuated the expression of IL-6 mRNA in HUVEC exposed to anti-β2GPI antibodies or cytokines. Incubation of HUVECs with mevalonate (100 μM), concomitantly with fluvastatin, greatly prevented the inhibitory effect of statin.
Endothelial activation mediated by anti-β2GPI antibody can be inhibited by statins. Because of the suggested role of endothelial cell activation in the pathogenesis of antiphospholipid syndrome (APS), our data provide, for the first time, a rationale for using statins as an additional therapeutic tool in APS.
The antiphospholipid syndrome (APS) is a recently described autoimmune disease characterized by the presence of elevated antiphospholipid antibodies (aPL) accompanied by the occurrence of recurrent arterial/venous thrombosis and fetal loss (1). The aPL appear to play a pathogenic role, rather than just being a simple diagnostic marker of the syndrome (2, 3). They constitute a heterogeneous family of autoantibodies with different specificities that are mainly directed against phospholipid-binding plasma proteins, of which β2-glycoprotein I (β2GPI) is now believed to represent the most important antigen target for aPL (2–7).
In both in vitro and in vivo models, aPL can react with β2GPI expressed on the endothelial cell surface, and this results in a proadhesive and a proinflammatory endothelial phenotype (7–9). The antibody-mediated crosslinking of β2GPI has been suggested to play an important role in inducing a cell signal that ends in adhesion molecule expression, cytokine secretion, accelerated prostacyclin metabolism, and tissue factor (TF) synthesis (10–12). Endothelial activation itself was suggested to be crucial in inducing the thrombophilic diathesis of APS (3, 8, 9).
Early atherosclerosis is now becoming recognized as one of the most serious complications in the management of patients with systemic autoimmune diseases, and a close link between atheroma and aPL recently has been pointed out (13, 14). The hypothesis that aPL, besides their ability to mediate a thrombophilic diathesis and an adverse pregnancy outcome, may have a potential role in the accelerated arterial disease observed in APS is related to their ability to induce endothelial activation, which could be relevant in the onset of the atherosclerotic process (15). Additional findings appear to support such a hypothesis: β2GPI has been shown to be localized in atherosclerotic plaques (16), and mice lacking the low-density lipoprotein (LDL) receptor develop accelerated atherosclerosis after active immunization with β2GPI (17). Moreover, in vitro studies have shown that aPL cross-react with oxidized LDL (ox-LDL) (18), and that they can even enhance the uptake of ox-LDL by macrophages (19).
Statins are cholesterol-lowering drugs that act by a competitive inhibition of the hydroxymethylglutaryl–coenzyme A (HMG-CoA) reductase enzyme (20). Large clinical studies have shown that, besides its cholesterol-lowering activity, statin treatment is associated with regression of atherosclerotic lesions and with a reduction of cardiovascular complications (21–24). Experimental evidence suggests that statins may influence several events in the vessel wall that are relevant for the progression of atherosclerosis, such as smooth muscle cell migration and proliferation (25), monocyte activation in terms of cytokine and metalloprotease synthesis (26–28), and TF expression (29). More interestingly, statins can also affect endothelial cell function by preventing the reduction of endothelial cell nitric oxide synthase (NOS) induced by ox-LDL (30), by increasing tissue-type plasminogen activator synthesis (31), by inhibiting TF expression after cytokine or lipopolysaccharide (LPS) activation (32), and by promoting angiogenesis (33). All these effects appear to be primarily related to HMG-CoA reductase inhibition, since intermediates of the mevalonate pathway might reverse such an effect (27, 31–33).
The aim of the present study was to investigate whether statins are able to affect endothelial activation induced by anti-β2GPI antibodies and to elucidate the mechanism by which the drugs are active. From our results, it appears that statins might inhibit the anti-β2-GPI–induced proadhesive and proinflammatory endothelial phenotype, and that such an inhibition occurs at the nuclear level. Taken together, these findings suggest the potential use of statins as an additional tool for the treatment of the clinical manifestations related to aPL.
Fluvastatin was a generous gift from Novartis Pharma (Basel, Switzerland), while simvastatin was kindly provided by Merck, Sharp, and Dohme (Rome, Italy). Mevalonate was obtained from Sigma (St. Louis, MO). Simvastatin in lactone form was brought into solution by 0.1M NaOH to give the active form and the pH was adjusted to 7.4.
Recombinant human interleukin-1β (rHuIL-1β) and tumor necrosis factor α (TNFα) were from R&D Systems (Minneapolis, MN), while the LPS was from Sigma. The monoclonal anti-human adhesion molecules (E-selectin and intercellular adhesion molecule 1 [ICAM-1]) were from R&D Systems.
Human umbilical vein endothelial cells (HUVECs) were cultured in medium 199 (E199; Flow, Irvine, UK) supplemented by 20% fetal calf serum (FCS; ICN Biomedicals, Costa Mesa, CA) and 100 μM of penicillin-streptomycin, as described (10). We previously demonstrated that bovine β2GPI supplied by FCS in the medium can adhere to the HUVEC cell membrane and can be recognized by human polyclonal and monoclonal anti-β2GPI antibodies (10, 11).
HUVECs were isolated from normal-term umbilical cord veins by collagenase perfusion and cultured as previously reported (10).
Two human IgM monoclonal antibodies (mAb) derived from patients with APS were used. GR1D5 has been previously characterized as reacting with human β2GPI, and TM1B9, which did not display any reactivity to β2GPI, was used as a negative control. The characterization of these mAb has been previously reported in detail (34, 35).
Anti-β2GPI antibodies were detected by a solid-phase enzyme-linked immunosorbent assay (ELISA), as described (10, 35–38). Sera from 2 patients with primary APS were collected and the whole IgG fractions were purified on protein G–Sepharose (HiTrap Protein G; Pharmacia Biotech, Freiburg, Germany) as previously reported (36). The anti-β2GPI antibodies were subsequently affinity purified using β2GPI columns as previously described (36). The affinity-purified IgG reacted specifically with β2GPI as detailed previously (36, 38).
Endothelial adhesion molecule expression was evaluated by a cell ELISA, as previously described (10). The cells were incubated in complete medium overnight with different concentrations of fluvastatin (1–10 μM), fluvastatin (5 μM) plus mevalonate (100 μM), simvastatin (5 μM), or simvastatin (5 μM) plus mevalonate (100 μM). Expression of E-selectin and ICAM-1 was induced by adding to the medium 1) rHuTNFα (10 ng/ml), 2) rHuIL-1β (50 units/ml), 3) LPS (20 ng/ml), 4) human IgM anti-β2GPI mAb (GR1D5; 100 μg/ml), or 5) affinity-purified human IgG anti-β2GPI (100 μg/ml). For negative controls, the cells were incubated with 1) human irrelevant IgM mAb (TM1B9; 100 μg/ml), 2) normal human serum (NHS) IgG (100 μg/ml), or 3) medium alone. Cell viability was >90% when evaluated by trypan blue exclusion, and no cell loss could be detected by an optical density (OD) cell-counting method previously described (38) (data not shown).
The assay was performed with 51CrNa (30 μCi/106 cells; Amersham, Little Chalfont, UK) labeled U937 cells as previously described (39). Adhesion assays were performed on HUVEC monolayers pretreated overnight with fluvastatin (5 μM) or fluvastatin (5 μM) plus mevalonate (100 μM). HUVEC monolayers were activated by 100 μg/ml of the mAb or polyclonal anti–β2GPI antibodies, or by TNFα (10 ng/ml), IL-1β (50 units/ml), or LPS (20 ng/ml) for 5 hours. The same negative controls mentioned above were used. Cell viability was >90% when evaluated by trypan blue exclusion, and no cell loss could be detected by the OD cell-counting method previously described (38) (data not shown).
Nuclear extract preparation. HUVEC monolayers were incubated overnight with fluvastatin (5 μM) and then stimulated with TNFα (10 ng/ml) or with positive and negative anti-β2GPI antibodies (100 μg/ml) for 30 minutes. For nuclear protein extraction, the cells were harvested in ice-cold Tris buffered saline/KCl buffered saline (25 mM Tris base, 137 mM NaCl, 2.7 mM KCl; pH 7.4), and centrifuged at 4,000 revolutions per minute for 30 seconds. The pellet was resuspended in 100 μl buffer A (10 mM HEPES, pH 7.9; 1.5 mM MgCl2; 10 mM KCl; 0.5 mM dithiothreitol [DTT]) and incubated on ice for 15 minutes. After the addition of 4 μl 10% Triton X-100, samples were vortexed for 10 seconds and centrifuged at 13,000 rpm for 1 minute at 4°C. Pellets were resuspended in 50 μl of cold buffer B (20 mM HEPES, pH 7.9; 1.5 mM MgCl2; 25% glycerol; 0.4M NaCl; 0.2 mM EDTA; 0.5 mM phenylmethylsulfonyl fluoride; 0.5 mM DTT) and incubated on ice for 30 minutes. After centrifugation (13,000 rpm for 10 minutes at 4°C), the supernatants were frozen (−80°C). The protein concentration of the nuclear extract was measured by the Lowry method (40).
The oligonucleotide containing the sequence of the E-selectin NF-κB site (5′-TGGATATTCCCGGGAAAGTTTTTGGAT-3′), obtained from Oligos Etc. (Wilsonville, OR), was annealed with a complementary primer and radiolabeled with α-32P-dCTP (Amersham) by the random priming technique and separated from unincorporated nucleotide using a Sephadex G-25 (Pharmacia Biotech) spun column. Gel-shift assays were performed as follows: 4 μg of nuclear extracts were incubated with radiolabeled DNA probes (1 × 105 counts per minute) for 20 minutes at room temperature in 18 μl binding buffer (60 mM KCl, 1 mM EDTA, 20 mM HEPES, 12% glycerol) containing 15 mM DTT and 4 μg dI-dC. Protein–DNA complexes were separated from the free DNA probe by electrophoresis through 5% nondenaturing acrylamide gels in 0.5× Tris–borate–EDTA (45 mM Tris base, 45 mM boric acid, 1 mM EDTA). At the end of the electrophoresis, gels were dried under vacuum on filter paper and exposed for 16–40 hours to Kodak XAR film (Eastman Kodak, Rochester, NY) at −80°C. Cold competition controls were performed by preincubating the nuclear proteins with unlabeled 100-fold molar excess of double-stranded oligonucleotides for 10 minutes, before the addition of the 32P-labeled oligonucleotide (41).
HUVECs were incubated overnight with fluvastatin (5 μM) or with medium alone and then stimulated for 4 hours, 30 minutes with TNFα (10 ng/ml) or with anti-β2GPI antibodies (100 μg/ml); NHS IgG (100 μg/ml) or medium alone served as negative controls. Total cellular RNA was obtained according to the method of Chomczynski and Sacchi (42). IL-6 and GAPDH messenger RNA (mRNA) expression was detected using the RiboQuant in vitro transcription kit according to the manufacturer's instructions (PharMingen, San Diego, CA).
Values were calculated as the mean ± SD. Statistical analyses were performed using one-way analysis of variance for multiple comparisons. P values less than 0.05 were considered significant.
Fluvastatin inhibits U937 adhesion to HUVEC. Endothelial activation by human affinity-purified polyclonal anti-β2GPI IgG (100 μg/ml) increased U937 adhesion to HUVEC monolayers (71 ± 4%, 77 ± 3%, and 14 ± 2%, mean ± SD percentage of adherent U937 for affinity-purified anti-β2GPI IgG samples 1 and 2 and NHS IgG, respectively), as previously reported (39). The human anti-β2GPI IgM mAb (GR1D5) gave comparable increases in U937 adhesion (data not shown). This rules out the possibility that the engagement of the Fcγ receptor on U937 was responsible for the increased adhesion to HUVEC. LPS (20 ng/ml), TNFα (10 ng/ml), and IL-1β (50 units/ml) exerted similar effects (81 ± 3%, 79 ± 4%, and 84 ± 2%, respectively).
Preincubation of HUVECs for 16 hours with fluvastatin (1–10 μM) reduced, in a concentration-dependent manner, the adhesion of U937 induced by polyclonal anti-β2GPI antibodies, with maximal effect at the 5-μM concentration (Figure 1A). Similarly, fluvastatin reduced, in a concentration-dependent manner, the adhesion of U937 to HUVEC activated by LPS, IL-1β, and TNFα (Figure 1A). Mevalonate (100 μM) prevented the effects of fluvastatin (Figure 1B), which suggests that fluvastatin inhibits monocyte adhesion to HUVECs with a mechanism involving the inhibition of the HMG-CoA reductase enzyme.
Statins inhibit E-selectin and ICAM-1 expression in HUVECs. To explore the mechanisms responsible for the inhibition of monocyte adhesion exerted by fluvastatin, we investigated its effects on E-selectin and ICAM-1 expression by HUVECs exposed to the different stimuli. Incubation of HUVECs with human affinity-purified polyclonal IgG anti-β2GPI antibodies increased E-selectin and ICAM-1 expression, and these effects were comparable with those exerted by IL-1β, TNFα, or LPS. Similar results were found when HUVECs were activated by the human anti-β2GPI IgM mAb (GR1D5) (data not shown); adhesion molecule expression was not influenced by polyclonal NHS IgG (Figures 2A and 3) or by the irrelevant human IgM mAb (data not shown).
Preincubation of HUVECs with fluvastatin reduced, in a concentration-dependent manner, the expression of E-selectin induced by anti-β2GPI antibodies, with a maximal effect (50% reduction) at the 5-μM concentration (Figure 2B). Similarly, the drug reduced the up-regulation of E-selectin induced by 50 units/ml IL-1β, 10 ng/ml TNFα, or 20 ng/ml LPS (Figure 2B). Mevalonate completely prevented the inhibition of E-selectin expression induced by the tested stimuli (Figure 2A).
Incubation of HUVECs with the above stimuli also increased the expression of ICAM-1. Fluvastatin at the 5-μM concentration reduced ICAM-1 expression by almost 50%. As for E-selectin expression, mevalonate prevented the effect of fluvastatin by >70% under all conditions tested (Figure 3). Similarly, simvastatin (5 μM), a semisynthetic statin with lipophilic properties, reduced E-selectin and ICAM-1 expression in HUVECs exposed to all the above stimuli. The effect of simvastatin, however, was lower compared with that of fluvastatin (Figures 4A and B).
Fluvastatin impairs NF-κB translocation. E-selectin gene expression is mainly regulated at the transcriptional level. Exposure of endothelial cells to inflammatory mediators, such as TNFα, IL-1β, or LPS, leads to a rapid increase in transcription rate through activation of the NF-κB transcription factor, which recognizes multiple regulatory elements in the E-selectin promoter region. We therefore evaluated whether fluvastatin affected the activation of this transcription factor in HUVECs. Unstimulated HUVECs exhibited no NF-κB binding, whereas binding of the E-selectin–specific NF-κB was observed in response to human anti-β2GPI polyclonal IgG or TNFα (Figure 5). Fluvastatin (5 μM) prevented NF-κB translocation. Similar results were obtained by using human anti-β2GPI mAb to activate HUVEC monolayers (data not shown). Competition experiments performed with a 100-fold excess of unlabeled κB oligonucleotide demonstrated the specificity of DNA binding activity.
Fluvastatin inhibits IL-6 mRNA expression in HUVECs. IL-6 mRNA expression was evaluated by RNA protection assay experiments. Human affinity-purified polyclonal IgG antibodies and TNFα up-regulated IL-6 mRNA expression in HUVECs (Figure 6). Human anti-β2GPI mAb GR1D5, but not the irrelevant control TM1B9, gave comparable results (data not shown). Fluvastatin markedly reduced the IL-6 mRNA expression under all conditions tested (Figure 6).
This study shows for the first time that fluvastatin and simvastatin prevent endothelial cell adhesive properties induced by anti-β2GPI antibodies and that this effect is a consequence of inhibition of NF-κB binding to DNA. The β2GPI-dependent aPL are widely accepted to play a pathogenic role in APS and have been shown to activate endothelial cells in different experimental models, both in vitro and in vivo (8, 9). Up-regulation of endothelial adhesion molecule expression has been particularly closely related to a procoagulant state in in vivo models (9). It is also worth mentioning that aPL detected by the anti-β2GPI assay appear to display a higher diagnostic specificity than the antibodies detected by the anticardiolipin assay (43). In this regard, the antibody preparations used in this study are representative of the pathogenic autoantibodies spontaneously occurring in the syndrome, further stressing the relevance of our data (34–36, 44).
We previously reported that anti-β2GPI antibodies increase cytokine secretion by HUVECs (10). We now show that the same antibodies also increase the expression of IL-6 mRNA in HUVECs and that the effect is prevented by fluvastatin. Thus, statins are capable of controlling both expression of adhesion molecules and cytokine production by vascular endothelium activated by anti-β2GPI antibodies. Such a protective effect of statins on the endothelium might consequently prevent monocyte adhesion/activation, a pathogenic mechanism relevant both in the induction of a procoagulant phenotype and in the early steps of the atherosclerotic process.
Interestingly, statins also prevent endothelial cell activation induced by cytokines or bacterial LPS. This observation is consistent with results of several studies that show that both fluvastatin and simvastatin can down-regulate endothelial adhesion molecule expression in different experimental models (45–48), but it is in contrast with an apparent potentiation reported by Sadeghi et al in which lower concentrations of the drug and submaximal amounts of IL-1β and TNFα were used (49). As a whole, these findings suggest that statins affect mechanisms common to different agonists, which operate downstream of the interaction of anti-β2GPI antibodies with the endothelial cells. Accordingly, fluvastatin impairs NF-κB translocation induced by TNFα as well as by monoclonal or polyclonal anti-β2GPI antibodies. We have previously shown that statins reduce TF expression by influencing RelC/p65 binding to DNA in human macrophages (32). Activation of the NF-κB/Rel transcription factor family by nuclear translocation of cytoplasmic complexes plays a central role in inflammation through its ability to induce transcription of proinflammatory genes. This pathway is activated upon appropriate cellular stimulation, most often by signals related to pathogens and stress (50).
In our study, anti-β2GPI antibodies are shown to induce this pathway, which, in turn, results in an activation of endothelial cells similar to that occurring with cytokines or bacterial LPS. Overexpression of NF-κB has been reported in diverse inflammatory diseases and, in particular, in inflammation associated with atherosclerosis (51). Thus, statins, through the control of inflammatory gene activation, may be effective in several conditions in which inflammatory endothelial activation is present. A newly discovered effect of statins as an effective repressor of major histocompatibility complex II expression has been reported, which suggests that statins may also act as immunomodulators (52). The high degree of patient tolerance of statins makes them a potentially welcome addition to the currently limited arsenal of immunosuppressive agents.
The activity of fluvastatin is reversed to a great extent by mevalonate, which suggests a link between the inhibition of HMG-CoA reductase enzyme and the activity of this drug. The inhibitory effects of statins reported here, however, are not dependent upon inhibition of cholesterol biosynthesis since, in our experimental conditions, exogenous cholesterol was provided to cultured cells. Thus, inhibition of endothelial proadhesive and proinflammatory phenotypes is another feature of these drugs, which have been previously reported to display a variety of pleiotropic effects independent of their capacity to lower cholesterol biosynthesis. These include inhibition of smooth muscle cell proliferation, TF, and matrix metalloproteinase expression in monocytes (53). In addition, statins have been reported to up-regulate plasminogen activator and constitutive NOS in endothelial cells (53). Some of these effects have been shown to be dependent upon inhibition of Rho family proteins. Interestingly, expression of Rho proteins has been shown to be required in endothelial cells for the assembly of stable adhesion with monocytes via the clustering of monocyte-binding receptors and their association with the actin cytoskeleton, independent of stress fiber formation (54).
Endothelial cell activation mediated by antiβ2GPI antibodies may result not only in the expression of an adhesive and inflammatory state but also in the development of procoagulant features, in part due to increased TF expression (12). Data reported here strongly suggest that statins may counteract endothelial activation induced by anti-β2GPI antibodies, at least in vitro; the finding that statin therapy may control both the proinflammatory and the procoagulant properties of vascular endothelium raises the possibility that it could be used in conjunction with standard therapy focused on the prevention of recurrences by oral anticoagulants and/or antiplatelet treatment. Interestingly, in our study, statins are effective at concentrations close to the levels occurring in vivo after their administration at therapeutic doses.
The endothelial activation mediated by antiβ2GPI antibodies has also been thought to have a potential role in the accelerated atherosclerosis associated with APS (14). These findings, together with the high degree of patient tolerance of statins, make them a potentially welcome addition to the current arsenal of pharmacologic agents in the treatment of the whole clinical spectrum of APS.