Statins, known as inhibitors of 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase, exhibit numerous functions related to inflammation, such as MHC class II down-regulation, interference with T cell adhesion, and induction of apoptosis. Here we demonstrate that both subcutaneous and oral administration of atorvastatin inhibit the development of actively induced chronic experimental autoimmune encephalomyelitis in SJL/J mice and significantly reduce the inflammatory infiltration into the central nervous system (CNS). When treatment was started after disease onset, atorvastatin reduced the incidence of relapses and protected from the development of further disability. Both the reduced autoreactive T cell response measured by proliferation toward the encephalitogenic peptide PLP139–151 and the cytokine profile indicate a potent blockade of T helper cell type 1 immune response. In in vitro assays atorvastatin not only inhibited antigen-specific responses, but also decreased T cell proliferation mediated by direct TCR engagement independently of MHC class II and LFA-1. Inhibition of proliferation was not due to apoptosis induction, but linked to a negative regulation on cell cycle progression. However, early T cell activation was unaffected, as reflected by unaltered calcium fluxes. Thus, our results provide evidence for a beneficial role of statins in the treatment of autoimmune attack on the CNS.
Several large-scale intervention trials have demonstrated that statin treatment reduces the risk of myocardial infarction and stroke (1). This effect was so far explained by their impact on the lipid profile through the inhibition of an enzyme crucial to cellular cholesterol synthesis, 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA)* reductase. However, increasing clinical and experimental evidence suggests that the pharmacological effects of statins relate not only to cholesterol lowering but also to anti-inflammatory effects on the immune system. Two independent studies have shown that both pravastatin and lovastatin reduce serum concentrations of C-reactive protein, a marker of inflammation, by ∼15% (2, 3). In addition, clinical trials on transplant rejections indicate the therapeutic potential of statins as immunosuppressive agents. For example, pravastatin improves survival and lowers the incidence of acute rejection after heart transplantation (4). Concerning the underlying mechanisms, Kwak et al. demonstrated in an initial report that statins inhibit the IFN-γ–induced expression of MHC class II on most APCs, including B cells and macrophages (5). A recent paper showed that certain statins inhibit the LFA-1–dependent stimulation of T cells and that a lovastatin-based LFA-1 inhibitor reduces infiltration of neutrophils in the murine thioglycollate-induced peritonitis model (6). Members of the statin family were found to interfere with the LFA-1–intercellular adhesion molecule (ICAM)-1 interaction by binding to a novel regulatory site of integrins in a HMG-CoA–independent manner. These observations raise the question as to whether statin therapy may also be beneficial in other chronic inflammatory conditions and immune diseases (7).
In this report, we used the pharmacologically active statin atorvastatin to dissect the role of statins in brain inflammation and to explore their therapeutic potential for autoimmunity. Therefore, we applied atorvastatin in murine experimental autoimmune encephalomyelitis (EAE; reference 8), a well-characterized CD4+ T cell–mediated animal disease model of multiple sclerosis (MS; references 9, 10). As for its human counterpart, induction of EAE in susceptible strains of rodents leads to central nervous system (CNS) inflammation, demyelination, and axonal pathology (11). We demonstrate that the signs of relapsing–remitting EAE in SJL/J mice can be diminished by atorvastatin in vivo both in a preventive and therapeutic fashion. CNS inflammation associated with EAE was significantly reduced. Our data indicate that this therapeutic effect is caused by a down-regulation of Th1 immune response, as confirmed by the reduction of autoreactive T cell response and infiltration into the CNS. We further demonstrate an HMG-CoA-reductase–dependent interference of atorvastatin with the cell cycle in human antigen–specific T cells.
Materials And Methods
Treatment with Atorvastatin and Mevalonate.
Because pure atorvastatin (Pfizer) is not soluble in PBS, the stock was dissolved in 2% DMSO before subcutaneous administration at a daily dose of 20 or 200 μg per mouse (corresponding to 1 or 10 mg/kg body weight) and in methylcellulose 0.5% (Sigma-Aldrich) before oral administration of 200 μg per mouse (10 mg/kg) for the in vivo experiments. Accordingly, the in vitro investigations were performed at a concentration of 1, 5, and 25 μM of atorvastatin (100 μM stock in 2% DMSO). In all assays of our work, the carrier was used as vehicle control in the same dilution as atorvastatin. To test the reversibility of the atorvastatin effect, l-mevalonate was used at a concentration of 200 μM. Therefore, l-mevalonic acid lactone (Sigma-Aldrich) was activated by 1 N NaOH. The resulting solution was neutralized with 1 N HCl to pH 7.2, diluted with distilled water, and sterilized by filtration thereafter.
Induction, Treatment, and Clinical Evaluation of Active EAE.
Female SJL/J mice (6–8 wk; ∼20 g body weight; Charles River Laboratories) were immunized subcutaneously with 75 μg proteolipid protein (PLP) 139–151 (purity >95%; Pepceuticals) in 0.2 ml emulsion consisting of equal volumes of PBS and CFA (Difco) and containing 6 mg/ml of mycobacterium tuberculosis H37Ra (Difco). 200 ng pertussis toxin (List Biological Laboratories) was administered intraperitoneally at days 0 and 2 (12). Mice were scored for EAE as follows: 0, no disease; 1, tail weakness; 2, paraparesis; 3, paraplegia; 4, paraplegia with forelimb weakness or paralysis; and 5, moribund or dead animals (13). Mean clinical scores at separate days and mean maximal scores were calculated by adding scores of individual mice and dividing by number of mice in each group. All procedures were conducted according to protocols approved by the local animal welfare committee.
Mice killed with narcotics were transcardially perfused with 4% paraformaldehyde. Spinal cords and brain stems were removed, postfixed in 4% paraformaldehyde, and embedded in paraffin. Vibratome sections (Shandon) were stained with hematoxylin and eosin for visualization of inflammatory infiltrates and evaluated in a blinded manner for the amount of inflammation as described previously (14).
Human and Murine Antigen-specific T Cells.
Human CD4+ T cell lines (TCLs) specific for human myelin basic protein (MBP; MA14) or birch pollen (BP; MB7, MB8, EG8) were established with a modified “split-well” protocol (15) and have been characterized previously (16). All antigen-specific TCLs had at least three restimulations before the assay and a stimulation index >3. Establishment of the PLP139–151-specific mouse TCL Je1 was performed as described previously (17). For murine bulk cell cultures, spleen and draining lymph node cells were dissociated with a cell strainer (Sigma-Aldrich) and cultured in 96- (for proliferation assays) or 24-well microtiter plates (for cytokine assays).
Proliferation and Cytokine Assays.
For the proliferation of human TCLs, 0.7 × 105 cells were stimulated with 2 × 105 autologous-irradiated (3,000 rad) PBMC as antigen-presenting cells and the corresponding antigen, or with anti-CD3 (OKT3; coated at 1 μg/ml; American Type Culture Collection) and 1 μg/ml anti-CD28 (R&D Systems). For the proliferation of the mouse PLP139–151-specific TCL Je1, 0.7 × 105 cells were stimulated with 1.4 × 105 autologous-irradiated (3,000 rad) splenocytes as APC and 5 μg/ml peptide. For proliferation assays, [3H]Thymidine (0.5 μCi per well; Amersham Biosciences) incorporation was measured 72 h after stimulation with antigen and 48 h with anti-CD3/CD28, radioactivity as counts per minutes detected by a MicroBeta® β counter (PerkinElmer) after 18 h. For the ex vivo proliferation and cytokine measurement, 2 × 106 splenocytes were stimulated with different peptide concentrations. Human IFN-γ and mouse IL-2, IL-12, IFN-γ, IL-4, and IL-10 were measured in supernatants from cell cultures (2 × 106/ml in 24-well plates; incubation time 48 h for IL-2, IL-12, and IFN-γ; 96 h for IL-4 and IL-10) with commercially available sandwich ELISA® kits (mouse, BD Biosciences; human, R&D Systems).
Immunoblot studies were performed as described previously (16). In brief, after 2-h blocking, membranes were incubated for 1 h with specific primary antibodies: 2 μg/ml monoclonal mouse anti-p27Kip1 (Santa Cruz Biotechnology, Inc.), 1 μg/ml polyclonal rabbit anti-CDK4 (Santa Cruz Biotechnology, Inc.), and 24 ng/ml monoclonal anti–β-actin. Membranes were incubated for 1 h with 1.25 μg/ml secondary antibody coupled to horseradish peroxidase (DakoCytomation). Specific bands were detected with the ECL-Plus™ System (Amersham Biosciences). CDK4 and p27Kip1 immunoblots were sequentially incubated with anti–β-actin as control.
Measurements of Intracellular Calcium [Ca2+]i.
Intracellular free calcium concentrations [Ca2+]i were measured in human TCLs using flow cytometry (18). Cells, 5 × 106/ml, were loaded with 5 μM Indo-1-AM (Molecular Probes) for 30 min at 37°C. Unloaded dye was removed by centrifugation, and cells were resuspended in calcium-free PBS. The analyses were performed on a flow-activated cell sorter (BD FACS® LSR; Becton Dickinson). The FL4 (510/20) and FL5 (380LP) fluorescence channels were used to measure free Indo-1 and the complex Ca2+–Indo-1 concentrations, respectively. Calculation of the ratio of these two fluorescent wavelengths allows the evaluation of changes in cytosolic-free Ca2+ concentrations [Ca2+]i independently of the cell size and the intracellular Indo-1 concentration. The flow rate was set to 300 events/s and the mean ratio of 2,000 cells were noted every 20 s. 5 μM Tg (Calbiochem) was added after 80 s to fully deplete intracellular Ca2+ stores and therefore activate calcium release-activated calcium channels (CRACs). After a 5-min incubation with Tg, 1.2 mM calcium was added to the cell suspension to monitor the extent of Ca2+ influx.
DNA fragmentation of apoptosis-susceptible Jurkat T cells was determined as described previously (15). 2 × 105 cells were incubated with atorvastatin or vehicle at the indicated concentrations. As a positive control, CD95-mediated apoptosis was induced by a soluble murine CD95 ligand. After 24 h, cells were washed and incubated overnight at 4°C in a hypotonic fluorochrome solution containing 50 μg/ml propidium iodide before flow cytometry for quantitative detection of hypodiploid DNA.
For group comparisons (mean maximal disease scores and number of inflammatory foci in spinal cord), the Student's t test was applied. To compare the EAE courses, statistical analysis of repeated measurements was performed by analysis of variance using the Fisher's pairwise least significant difference test with Bonferroni adjustment (19). The incidence of relapses (defined as the acquisition of a clinical score >1 for at least two consecutive days; modified according to reference 20) was compared using the χ2 test.
Results And Discussion
Atorvastatin Ameliorates Disease Severity in EAE.
To explore the efficacy of statins on actively induced EAE, PLP139–151-immunized SJL/J mice were treated subcutaneously with 200 μg per mouse (corresponding to 10 mg/kg body weight) atorvastatin (all treatments using pure substance) from day 1 after immunization. This dosage was chosen in accordance with previous work on statin effects in murine lipid metabolism (21–23) and stroke models (24, 25), assuming that mouse and man greatly differ in their statin metabolism (26–29). Preventive treatment attenuated the first disease attack and reduced clinical symptoms (Fig. 1 A), resulting in a reduced mean maximum disease score of 2.4 ± 0.3 (mean ± SEM) in the atorvastatin group versus 3.4 ± 0.3 in the vehicle group (data from two independent experiments; n = 12 for both groups; P < 0.05). Continued atorvastatin therapy of these animals significantly protected them from the development of long-term neurological sequelae, as the mean clinical scores between groups were statistically different (F14–70 [1, 57] = 4.4, P < 0.05). This marked prophylactic effect of atorvastatin led us to investigate its therapeutic potential to suppress established disease. Hence, we delayed starting treatment with atorvastatin until animals developed the first symptoms. Treatment with atorvastatin prevented mice from disease exacerbation with a relapse rate of 37.5% (three out of eight) in the atorvastatin group compared with 87.5% (seven out of eight) in the vehicle group (χ2 = 4.3; P < 0.05). Neurological disability during relapse was reduced (Fig. 1 B) with a maximum disease score of 1.2 ± 0.4 in the atorvastatin group versus 2.4 ± 0.4 in the vehicle group (P < 0.05), and significantly lower mean clinical disease scores (F35–50 [1, 15] = 4.7; P < 0.05). The lower dose of 20 μg per mouse atorvastatin was less effective (F18–20 [1, 3] = 5.1; P < 0.05; Fig. 1 C). Treatment with orally administered atorvastatin (200 μg per mouse) exhibited a beneficial effect in the same model (F25–28 [1, 4] = 4.8; P < 0.05; Fig. 1 D). Our observations provide evidence that the HMG-CoA-reductase inhibitor atorvastatin has marked therapeutic effects on chronic relapsing-remitting autoimmune encephalomyelitis in SJL/J mice. Clinical disease in atorvastatin-treated mice was considerably milder compared with vehicle-treated controls, as displayed by reduced maximal disease severity and favorable long-term outcome. Moreover, therapy with atorvastatin was beneficial even when started after disease manifestation, indicating a role for therapy of an already established autoimmune disorder.
Histological examination of mice killed directly after the maximum of the relapse revealed large inflammatory lesions scattered throughout the brain stem and spinal cord in vehicle-treated mice, whereas those treated with the statin revealed significantly less inflammation (Fig. 2). Thus, in line with the clinical scores, we found that atorvastatin treatment resulted in reduced CNS inflammation. Taking the reported almost pleiotropic effects of the statins into account, possible explanations for these data are either a decrease in peripheral immune response or reduced transmigration into the CNS. Therefore, we addressed the mechanisms underlying the atorvastatin-mediated blockade of neuroinflammation.
Atorvastatin Treatment Inhibits T Cell Response in EAE.
First, we examined the T cell response resulting from statin treatment in EAE, which was investigated after 10 d of treatment in cells from spleens exposed to the encephalitogenic PLP139–151 epitope. As shown in Fig. 3 A, mice treated with atorvastatin revealed a lower proliferative response after antigenic challenge, compared with vehicle-treated animals. In line with previous reports (25), treatment of normocholesterolemic SJL/J mice with atorvastatin for 10 d in contrast to therapy over several months did not alter serum cholesterol levels, with 57.3 ± 2.9 versus 55.7 ± 1.2 mg/dl in atorvastatin- versus vehicle-treated mice, respectively (mean ± SEM).
Inhibition of the Th1 Response by Atorvastatin in EAE.
Assessment of the cytokine pattern in response to PLP restimulation indicated that atorvastatin treatment led to a reduction of IL-2 and IL-12 as well as IFN-γ secretion, as demonstrated for splenocyte cultures (Fig. 3, B–D). Interestingly, although immune cells from atorvastatin-treated animals showed a diminished proliferative response toward PLP, the levels of IL-4 and IL-10 in the supernatants of splenocytes were even higher on atorvastatin treatment, as compared with vehicle-treated animals (Fig. 3, E and F). IL-12 drives the T cell differentiation toward the Th1, and IFN-γ is a proinflammatory cytokine, possibly involved in the initiation of exacerbations in relapsing-remitting MS (30). In EAE, IFN-γ directs trafficking of leukocytes into the CNS by orchestrating chemokine production (31), and its overproduction is usually associated with relapses (32). Thus, inhibition of proliferation, IL-12, and IFN-γ production points to an antiinflammatory activity of atorvastatin, which accounts for the reduction of clinical disease. Our results demonstrate that atorvastatin is an effective immunomodulatory agent for the treatment of murine EAE by suppressing disease-causing Th1 response while enhancing regulatory cytokines. Clinical trials may clarify whether the same effects can be achieved in MS patients.
Decreased Proliferation of Murine and Human Lymphocytes Associated with Interference in Cell Cycle Regulation.
In view of the reduced proliferative response in atorvastatin-treated animals, we investigated whether this statin interferes with the antigen-specific proliferation of a murine CD4+ Th1 PLP-specific TCL Je1 in vitro. As shown in Fig. 4 A, atorvastatin blocked the antigen-response in a dose-dependent manner. The doses of 1–25 µM atorvastatin used in these assays are comparable to the levels measured in human plasma (29, 33). Having demonstrated the inhibitory effects of atorvastatin on murine cells in vivo and in vitro, we next studied the proliferation of human antigen-specific CD4+ TCLs. As indicated in Fig. 4 B, antigen-specific proliferation of the human TCL MB8 was suppressed by atorvastatin in a dose-dependent manner.
Interestingly, the blockade of proliferation was also observed when atorvastatin was added to proliferating T cells, 24 or 48 h after stimulation (Fig. 5 A). This points to further pathways of immunomodulatory statin function in addition to those involving a reduced MHC class II up-regulation by inhibition of the inducible promoter IV of the transactivator CIITA (5), and the blockade of LFA-1–ICAM-1 interactions (6). To confirm this hypothesis, we tested the ability of atorvastatin to block T cell proliferation in response to direct T cell receptor engagement independent of antigen presentation. As shown in Fig. 5 B, [3H]Thymidine uptake of the antigen-specific TCL MB7 stimulated with anti-CD3/CD28 was markedly suppressed by atorvastatin. Thus, in an environment lacking APC, expressing MHC class II and ICAM-1, atorvastatin is nonetheless capable of inhibiting proliferation. Costimulation via T cell–mounted LFA-1 needs the addition of ICAM-1 not expressed on the T cell itself because direct T cell stimulation via CD3–TCR pathway alone does not induce phosphatidylinositol 3-kinase activation, a reliable indicator of costimulation mediated through ICAM-1 into the T cell (34). Therefore, a down-regulation of LFA-1 expression resulting in attenuated transmigration of mononuclear cells into the CNS (35), as well as ligation to the novel LFA-1 binding site as demonstrated by Weitz-Schmidt et al. (6), is not sufficient to explain the antiproliferative effect of atorvastatin in our CD3/CD28-stimulated T cells.
Reversibility of Atorvastatin-induced Interference with T Cell Proliferation and IFN-γ Secretion by l-Mevalonate.
Because atorvastatin may confer its immunomodulatory effects, both dependent on (5) and independent of (6) inhibition of HMG-CoA reductase, it was important to clarify whether the direct T cell–targeted antiproliferative effect reported here is mediated by the HMG-CoA reductase pathway. Indeed, both proliferation (Fig. 5 B) and IFN-γ secretion (Fig. 5 C) in human antigen–specific T cells could be reversed by l-mevalonate providing direct evidence that the immunomodulatory effects of atorvastatin are mediated by inhibition of HMG-CoA reductase.
Impact of Atorvastatin on Cell Cycle Regulation.
Because statins block the synthesis of isoprenoid intermediates and subsequently inhibit the function of intracellular signaling molecules (36), next we tested whether the observed antiproliferative effect of atorvastatin is caused by interference at the cell cycle level. We did not find an impact of atorvastatin on early T cell activation, because calcium influx was unaffected (Fig. 5 D). However, we could observe that growth inhibition by atorvastatin was associated with a down-regulation of the cyclin-dependent kinase 4 (CDK4; Fig. 5 E). Investigating the expression of p27kip1, a CDK inhibitor involved in negative regulation of the T cell cycle, we found an up-regulation on atorvastatin incubation, which could be reversed by l-mevalonate (Fig. 5 F). This finding, together with a persisting effect of atorvastatin, even when added late in culture (Fig. 5 A), suggests an inhibition of cell cycle progression at the late part of the G1 phase.
No Induction of Apoptosis in T Cells by Atorvastatin.
Principally, atorvastatin may not only inhibit but also directly eliminate cells by induction of apoptotic mechanisms (37). Testing statin-incubated T cells with the trypan blue exclusion assay, we had no indication of increased cell death. To exclude an apoptosis-inducing capacity of atorvastatin in T cells, which might be responsible for the antiproliferative effect, we measured DNA fragmentation. However, atorvastatin did not induce relevant DNA fragmentation, even in apoptosis-susceptible Jurkat T cells, compared with both untreated and vehicle-treated cells (Fig. 5 G), thus ruling out a proapoptotic role for this substance.
Disease pathogenesis in EAE induced by PLP139–151 in SJL/J mice is regarded as a prototypical T cell–mediated autoimmune disorder. As in MS, it involves the activation and expansion of autoreactive T cells, migration through the blood–brain barrier, restimulation by antigen challenge within the CNS, and induction of local inflammation, leading to demyelination and axonal damage. The reduced amount of inflammatory infiltration in the CNS from statin-treated mice might reflect the inhibition of T cell entry into the CNS. This hypothesis is supported by the fact that certain statins were capable of inhibiting thioglycollate-induced neutrophil migration (6). However, our data on the suppression of Th1 immune response outside the brain and the lack of the statin-binding site in the integrin very late antigen-4 (6), which is important for the transmigration through the blood–brain barrier, argue against a prominent impact of statins on T cell migration into the CNS. Thus, the reduced amount of inflammation in the CNS of statin-treated animals in our work most likely results from an effect on the peripheral immune response. In fact, in line with Youssef et al. (38), we found a diminished proliferative response of CD4+ lymphocytes toward the encephalitogenic peptide PLP139–151, accompanied by reduced production of Th1 cytokines and a relative increase in the regulatory cytokines IL-4 and IL-10. It is striking that even the addition of the statin at a late stage of T cell activation resulted in a marked decrease of proliferation. This finding indicates that the immunomodulatory effects of atorvastatin are comprised of other mechanisms in addition to down-regulation of MHC class II expression and inhibition of LFA-1–ICAM-1 interaction.
Investigations in the murine system indicate that statins exert pleiotropic immunomodulatory effects on both APC and T cells (38). In line with previous observations on lovastatin and B cells (39), the antiproliferative effect of atorvastatin examined in human antigen-specific T cells in our work was not preceded by a reduced T cell activation because calcium influx was unaffected. Furthermore, this effect was not linked to the induction of apoptosis. Our data indicate that the underlying mechanism for the inhibition of T cell response in autoimmune neuroinflammation is the interference with cell cycle regulation represented by down-regulation of CDK4 and up-regulation of p27kip1, which was reported previously as mechanism of action of statins only in mesangial cells (40). Immunological studies within clinical trials might finally clarify which atorvastatin-mediated mechanism is important in case of a beneficial effect in MS.
Via blocking HMG-CoA reductase and mevalonate synthesis, statins prevent the synthesis of important isoprenoid intermediates of the cholesterol biosynthetic pathway, which are important for isoprenylation of certain cell-signaling proteins (36). Thus, this posttranslational modification of intracellular signaling molecules, such as small GTP-binding proteins by isoprenylation (36), is involved in the regulation of the cell cycle by statins. Small GTP-binding proteins (e.g., Ras and Rho) require such posttranslational modification for membrane localization and activity and are implicated in cell cycle regulation. Among these proteins, Ras promotes cell cycle progression via activation of the mitogen-activated protein kinase pathway (41), whereas Rho causes cellular proliferation possibly through destabilizing p27kip1 protein (42). In fact, we could show that atorvastatin-induced up-regulation of p27kip1 is blocked by l-mevalonate. Altogether, this pathway explains the down-regulation of CDK4 and up-regulation of p27kip1 by atorvastatin.
Our data provide compelling evidence for the beneficial role of statins in neuroinflammation. In addition to the reported pleiotropic mechanisms exhibited by this group of substances, here we unravel the inhibition of HMG-CoA reductase as a responsible mechanism for the direct T cell–targeted influence of atorvastatin on proliferation and effector cytokines of Th1 cells. Due to the therapeutic potential of this pathway in EAE, the activity of statins might open up a new avenue for MS treatment with an orally administered drug.
We thank Michael Rupschus for technical assistance, Prof. Steffen Hauptmann (Charité) for technical support, Dr. Reinhard Geßner (Charité) for cholesterol analysis, Dr. Toralf Kaiser (Joint Flow Cytometry Core Facility of the Max-Planck-Institute for Infection Biology, German Rheumatism Research Center, and Charité) for assistance with calcium measurements, Dr. Julian Bösel and Dr. Karen Gertz for technical hints, and Celia Forbes for carefully reading the manuscript as a native speaker.
This work was supported by grants from the Bundesministerium für Bildung und Forschung to F. Zipp and S. Brocke, the Deutsche Forschungsgemeinschaft to F. Zipp and R. Nitsch (507), and to M. Endres (En343/7-1; 4-2; 6-2), the Hermann and Lilly-Schilling Foundation to M. Endres, and the Boehringer Ingelheim Fonds to O. Aktas.
Abbreviations used in this paper: BP, birch pollen; CDK, cyclin-dependent kinase; CNS, central nervous system; EAE, encephalomyelitis; HMG-CoA, 3-hydroxy-3-methylglutaryl coenzyme A; ICAM-1, intercellular adhesion molecule-1; MBP, myelin basic protein; MS, multiple sclerosis; PLP, proteolipid protein; TCL, T cell line.