Atorvastatin inhibits the immediate-early response gene EGR1 and improves the functional pro of CD4+T-lymphocytes in acute coronary syndromes

Background- Adaptive immune-response is associated with a worse outcome in acute coronary syndromes. Statins have anti-inflammatory activity beyond lowering lipid levels. We investigated the effects of ex-vivo and in-vivo atorvastatin treatment in acute coronary syndromes on CD4+T-cells, and the underlying molecular mechanisms. Approach and results- Blood samples were collected from 50 statin-naïve acute coronary syndrome patients. We assessed CD4+T-cell activation by flow-cytometry, the expression of 84 T-helper transcription-factors and 84 T-cell related genes by RT-qPCR, and protein expression by Western-blot, before and after 24-hours incubation with increasing doses of atorvastatin: 3-10-26 g/ml (corresponding to blood levels achieved with doses of 10-40-80 mg, respectively). After incubation, we found a significant decrease in interferon-?-producing CD4+CD28nullT-cells (P = 0.009) and a significant increase in interleukin-10-producing CD4+CD25highT-cells (P < 0.001). Atorvastatin increased the expression of 2 genes and decreased the expression of 12 genes (in particular, EGR1, FOS,CCR2 and toll like receptor-4; >3-fold changes). The in-vivo effects of atorvastatin were analyzed in 10 statin-free acute coronary syndrome patients at baseline, and after 24h and 48h of atorvastatin therapy (80 mg/daily): EGR1-gene expression decreased at 24h (P = 0.01) and 48h (P = 0.005); EGR1-protein levels decreased at 48h (P = 0.03). Conclusions-In acute coronary syndromes, the effects of atorvastatin on immune system might be partially related to the inhibition of the master regulator gene EGR1. Our finding might offer a causal explanation on why statins improve the early outcome in acute coronary syndromes.


INTRODUCTION
Over the past few years our understanding of the importance of inflammation in coronary instability is considerably increased [1].
CD4 + CD28 null T-cells are a subset of long-lived directly cytotoxic CD4 + T-lymphocytes producing large amount of the pro-inflammatory cytokine interferon-γ (IFN-γ), that have been implicated in the pathogenesis of various chronic inflammatory diseases. We previously demonstrated that circulating CD4 + CD28 null T-cell frequency higher than 4% increase the risk of acute coronary syndromes (ACS), particularly in diabetic patients [2,3].
At the other extreme, naturally occurring regulatory T-cells are a major cellular source of interleukin (IL)-10, a potent anti-inflammatory cytokine. These regulatory T-cells are involved in the control of autoimmunity [4].

Research Paper: Pathology
Accordingly, a lower number or a decreased function of these cells has been found in patients suffering from lupus erythematous, type-1 diabetes, rheumatoid arthritis, and multiple sclerosis [5], as well as in patients with ACS [6,7]. Moreover, in ACS the production of proinflammatory cytokines is not adequately counterbalanced by anti-inflammatory cytokines, such as IL-10; these alterations have been related to a worse short-and longterm prognosis [8,9].We recently observed that a subset of ACS patients presents an alteration of the immune response, associated to a worse outcome and characterized by reduced regulatory T-cell response to effector T-cell expansion [10].
Statins have anti-inflammatory and immunesuppressive activity besides lowering lipids that may, at least partially, account for their outcome improvement in the setting of both acute and chronic ischemic heart disease [11,12,13]. In particular, statins attenuate T-cell activation and proliferation, inhibit pro-inflammatory cytokine secretion and enhance anti-inflammatory cytokine secretion [14,15,16]. Two observational retrospective studies of our group have shown that in ACS the use of statins was associated with reduced levels of CD4 + CD28 null T-cells [17,2]. In a small number of ACS patients, rosuvastatin treatment for 6 weeks induced CD4 + CD28 null T-cell apoptosis [18]. Recent studies have also suggested that statins may enhance regulatory T-cell responses [19,20,21].
In the present study, we sought to investigate the effects of increasing doses of atorvastatin on phenotype and function of different CD4 + T-cell subsets, obtained from 50 statin-naïve patients presenting with non-ST elevation (NSTE)-ACS and raised levels of CD4 + CD28 null T-cells. To explore the mechanisms by which atorvastatin might suppress the immune response in ACS, we analyzed by quantitative PCR array the expression of 84 transcription factors involved in the immune response and 84 genes related to the functional properties of different T-helper cell subsets.
Finally, we assessed the in-vivo effects of high-dose of atorvastatin (80 mg/daily) in ACS patients.

RESULTS
Patient selection and study design are presented in Figure-1. Table 1 summarizes the clinical characteristics of the study population.
The percentage of total CD4 + T-cells, CD4 + CD28 null Tcells, CD4 + CD25 high T-cells and CD4 + CD25 high T-cells expressing the transcription factor Foxp3 did not change significantly after ex-vivo treatment with increasing doses of atorvastatin for 24 hours (Figure 2).

Effects of atorvastatin on pro-inflammatory and anti-inflammatory cytokine concentrations
IL-10 concentration increased from a median of 1.5 pg/mL, range 0.7-39.9 (untreated blood) to 6.3 pg/mL (range 1-42.7) after incubation of whole blood samples, collected from an antecubital vein at the time of patient enrollment, with 26 μg/ml of atorvastatin (P for trend = 0.024) (Figure-4, Panel C). Although we used a highsensitivity ELISA kit, IFN-γ was detectable in few patients at baseline and resulted undetectable after atorvastatin treatment (data not shown).

Atorvastatin decreases the expression of key cellular pathways in ACS CD4 + T-cells
To identify mechanisms by which atorvastatin might have immune-suppressive effects in CD4+T-cell populations, we analyzed the gene expression of a focused panel of 84 transcription factors downstream of signaling from cytokines, chemokines, growth factors, androgens and Toll-Like receptors. Then, we performed a PCR array profiling the expression of 84 genes including cytokine genes representative of the three classes of helper T-cells, genes encoding transcriptional factors that regulate the expression of these cytokines, markers of CD4+T-cell activation and other genes involved in the adaptive immune responses. PCR array analysis was applied on pooled RNA samples.
The complete list of genes investigated by PCR arrays, and their different expression after atorvastatin treatment, is reported in Table-2 and 3. PCR array analysis revealed that the expression of 2 genes was increased while the expression of 12 genes was decreased (>3-fold changes) by ex-vivo treatment of freshly isolated CD4 + T-cells from ACS patients with a dose of 26 μg/ml atorvastatin for 24 hours compared Figure 2: Effects of atorvastatin on total CD4 + T-cells, CD4 + CD28 null T-cells, CD4 + CD25 high T-cells and CD4 + CD25 high Foxp3 + T-cells. Panel A. Frequencies of total CD4 + and of CD4 + CD28 null T-cells were determined by flow-cytometry. CD4 + T-cells were isolated from peripheral blood samples of 20 statin-naïve NST-ACS patients and incubated for 24 hours without and with increasing doses of atorvastatin. Data are presented as median and 95% CI. The percentage of both total CD4 + (indicated in green) and of CD4 + CD28 null T-cells (indicated in red) did not change significantly after treatment with atorvastatin (P for trend = 0.337 and 0.080, respectively). Panel B. Frequencies of CD4 + CD25 high T-cells and of CD4 + CD25 high T-cells expressing the transcription factor Foxp3 were determined as described in Panel A. Data are presented as median and 95% CI. The percentage of both total CD4 + CD25 high T-cells (indicated in light blue) and of CD4 + CD25 high Foxp3 + T-cells (indicated in dark blue) showed slight, but not statistically significant, changes after treatment with atorvastatin (P for trend = 0.052 and 0.064, respectively). Panel C. Correlation between CD4 + CD25 high T-cells and CD4 + CD25 high Foxp3 + T-cells. Frequencies of CD4 + CD25 high T-cells and of CD4 + CD25 high T-cells expressing the transcription factor Foxp3 were calculated as percentage of CD4 + CD25 + T-cell population. A significant correlation was observed among these T-cell subsets (R = 0.67; P < 0.001). Spearman rank correlation was performed on pooled data (untreated/treated with increased doses of atorvastatin). For each condition (untreated/treated with atorvastatin) a pool of RNA was constituted, starting with the same amount of RNA (500 ng) from each patient of Cohort 1 (see Methods). Data are presented as fold of changes in gene expression induced by atorvastatin treatment as compared with the gene expression in untreated CD4 + T-cells. Increased genes are indicated in red and decreased genes in green. A fold change in gene expression of ≥ 3 was taken as significant.  For each condition (untreated/treated with atorvastatin) a pool of RNA was constituted, starting with the same amount of RNA (500 ng) from each patient of Cohort 1 (see Methods). Data are presented as fold of changes in gene expression induced by atorvastatin treatment as compared with the gene expression in untreated CD4+T-cells. Increased genes are indicated in red and decreased genes in green. A fold change in gene expression of ≥ 3 was taken as significant.
with control (Figure-5). Among the transcription factors, atorvastatin decreased the expression of early growth response 1 (EGR1) and V-fos FBJ murine osteosarcoma viral oncogene homolog (FOS). Among the genes related to helper T-cell pathway, atorvastatin decreased the expression of chemokine (C-C motif) receptor-2 (CCR2), the Toll-like receptor-4 (TLR4), and the proinflammatory cytokines IL-6 and IL-18.
To validate the PCR array results for selected genes, RT-qPCR was performed using RNA of each single patient. This set of data confirmed that atorvastatin significantly decreased the expression of CCR2 and of TLR4 as well as the expression of the transcription factors EGR1, FOS (Figure-6).
We also explored whether atorvastatin treatment might decrease EGR1 protein expression levels, as EGR1 gene showed the highest inhibition at PCR array analysis and also because EGR1 is a transcription factors critically involved in the immune response. Western-blot assay confirmed that incubation with atorvastatin (26 μg/ml) resulted in a significant reduction of EGR1 protein levels (from 5.0±2.6 to1.9±0.97; P = 0.038) (Figure-7).

In vivo effects of Atorvastatin
The in-vivo effects of atorvastatin were analyzed in 10 statin-free ACS patients at baseline, and after 24 hours and 48 hours of atorvastatin therapy (80 mg/daily). A single high-dose of atorvastatin has early effects on EGR1 mRNA expression and, with a reasonable delay, on EGR1 protein levels. In all patients, EGR1 gene expression was reduced after 24h of atorvastin therapy, from a mean (± SEM) value of 26.7 ± 5.7 at baseline to 8.5±1.9 at 24 hours (P = 0.01 versus baseline) and to 5.9±2.1 at 48hours (P = 0.005 versus baseline). Accordingly, EGR1 protein levels were significantly reduced after 48h of atorvastatin treatment, from a mean (± SEM) value of 26.1±2.2 at baseline to 24.7±1.9 at 24 hours (P = 0.67) and to 18.8±1.2 at 48 hours (P = 0.03 versus baseline) (Figure 8).

DISCUSSION
In the present study, we observed that atorvastatin ex-vivo, at concentrations corresponding to blood levels achieved with 10-40-80 mg/die, and after short-time (24 hours of incubation), modified the inflammatory activity of T-lymphocytes, although it did not affect their count. Atorvastatin reduced the frequency of CD4 + CD28 null Tcells producing IFN-γ, while it increased the production of the anti-inflammatory cytokine IL-10 by CD4 + CD25 high Tcells.
We also explored the mechanisms through which atorvastatin might have immune-suppressive effects in ACS, analyzing the expression of relevant transcription factors and genes related to T-cell functional properties. We found that atorvastatin significantly reduced the expression of the chemokine receptor CCR2, involved    Tables 2 and 3. www.impactjournals.com/oncotarget in trans-endothelial migration of inflammatory cells, and of TLR4. Both CCR2 and TLR4 induce a series of downstream transcriptional factors, including FOS and EGR1, finally leading to the production of a large number of pro-inflammatory cytokines [22,23,24]. Furthermore, EGR1 reduces IL-10 expression at the post-transcriptional level [25].
Finally, we explored the in-vivo effects of atorvastatin, demonstrating that a single high-dose of atorvastatin had also in-vivo early effects on EGR1 mRNA expression and on EGR1 protein levels. Moreover, it reduced the frequency of CD4 + CD28 null T-cells producing IFN-γ, and increased the production of the antiinflammatory cytokine IL-10 by CD4 + CD25 high T-cells.

Effects of atorvastatin on CD4 + T-cell functions
In ACS, CD4 + CD28 null T-cells are increased in peripheral blood [26] and infiltrate unstable coronary plaques where they undergo clonal expansion [27], probably triggered by specific antigens [28].They produce pro-inflammatory cytokines, in particular IFN-γ [26], and express high levels of TRAIL, a member of the TNF family implicated in apoptosis of vascular smooth muscle cells [29]. By directly stimulating apoptosis of vascular smooth muscle cells or by activating macrophages to kill these ones trough IFN-γ production, CD4 + CD28 null T-cells could weaken the fibrous cap and destabilize angiogenic vessels, precipitating atherosclerotic plaque rupture [30].
Moreover, these T-cells spontaneously express the subunit β1 of the IL-12 receptor even in the absence of antigenic stimulation, and respond to direct IL-12 stimulation with an increased expression of chemokine receptors that promote the tissue homing of effector T-cells [31]. In contrast, the anti-inflammatory cytokine IL-10 contributes to the atheroprotective effects of regulatory T-cells [32], and the expression of regulatory T-cells co-localizes with IL-10 within the atherosclerotic plaques [33]. ACS patients have a skewed helper T-cell differentiation oriented towards the expansion of aggressive effector T-cells and the reduction of Treg number and function [34]. These helper T-cell abnormalities characterize a sizeable proportion of ACS patients [10]. In this subset of ACS patients, the immune response might contribute to plaque destabilization through multiple damaging pathways [35].
Short-term treatment of ACS patients with statins is associated with a reduction of inflammatory markers [36], and with a rapid reduction in the intracellular production of TNF-α and IFN-γ by T-cells in-vitro [37]. Rosuvastatin, fluvastatin, and pitavastatin in-vitro treatment inhibits CD4 + T-cell-induced endothelial cell apoptosis by suppressing T-cell activation and TRAIL expression upon activation [38]. In two previous observational reports we found that the use of statins was associated with reduced levels of CD4 + CD28 null T cells [17,2].
Furthermore, short-term treatment of ACS patients with statins is associated with a significant increase of regulatory T-cell inhibitory properties and a significant reduction of serum IFN-γ and increase of IL-10 [39,40]. Statins may enhance regulatory T-cell responses by promoting their chemokine-dependent recruitment into inflammatory sites [21] and/or their differentiation in the periphery [19].

Effects of atorvastatin on the expression of inflammatory genes and transcription factors involved in the immune response
In our study atorvastatin reduced TLR4 gene expression, a pattern-recognition receptor stimulated by several pathogen associated molecular patterns (PAMPs) and damage associated molecular patterns (DAMPs). TLR4 has been found in atherosclerotic lesions and at the site of plaque rupture in patients with MI [41]; its expression is increased in thrombi [42] and in circulating monocytes [43] from patients with ACS. TLR4 stimulation induces intracellular pathways converging on nuclear factor (NF)-kB and mitogen activated protein kinases (MAPK), with subsequent release of pro-inflammatory cytokines and expression of co-stimulatory molecules [23].
Atorvastatin also reduced the expression of the chemokine receptor CCR2, that is involved in CD4 + T-cell transendothelial migration and recruitment at the site of tissue damage and inflammation [23]. CCR2 intracellular pathway also converges on MAPK and, downstream, on the nuclear transcription factors FOS and EGR1, both implicated in the immune response [22]. Indeed, we observed a decrease of the expression of these two transcription factors by atorvastatin. In particular, EGR1 gene showed the highest inhibition ( Figure 10). EGR1 is the prototype of a family of zincfinger transcription factors, "immediate-early response proteins", that is rapidly and transiently induced by a broad spectrum of extracellular signals, including growth factors, cytokines, injurious stimuli and many physiologic stimuli [44]. EGR1 is involved in cell growth, cell differentiation and cell survival [45]. EGR1 is expressed in T-cells and promotes T-cell activation and development by transcriptional induction of key cytokines, such as IL-2 and TNF-α, and costimulatory molecules, such as CD40 ligand, after T-cell receptor (TCR) stimulation [46,22]. EGR1 also induces the transcriptional activation of T-bet, the master gene regulator of Th1 [47], a T-cell subsets involved in atherosclerosis progression and plaque destabilization [48].
Conversely, EGR1 reduces the expression of IL-10 at post-transcriptional level, by inducing the transcription of a microRNA (hsa-miR-106a) that degrades IL-10 mRNA [25]. EGR1 might also promote atherogenesis through the activation of inflammatory genes [49,50]. Thus, we could speculate that, by reducing the expression of the transcription factor EGR1 in CD4+ T-cells, atorvastatin treatment might have the final net effect of reducing pro-inflammatory cytokine secretion, chemokine and chemokine receptor synthesis. This hypothesis is strongly supported by the in-vivo observation that a single high-dose of atorvastatin has early effects on EGR1 mRNA expression and, with a reasonable delay, on EGR1 protein levels. Moreover, it modifies the inflammatory profile of T-lymphocytes by decreasing IFN-γ production by aggressive effector T-cells, and increasing the production of the anti-inflammatory cytokine IL-10 by T-cells with a regulatory phenotype (Figure 10).
Although some of these beneficial effects have already been attributed to different statins in humans [51,52] as well as in animal models [53] and in-vitro [54], our study investigates a comprehensive expression profile in ACS.
In the setting of ACS, it has been proposed that the early outcome improvement observed with intensive statin treatment, compared to a moderate treatment schedule, might be related to anti-inflammatory properties rather than to lipid-lowering effects. According to this, the improvement of short-term outcome associated to intensive statin treatment appeared to correlate with hs-CRP level reduction rather than with LDL-cholesterol  [55]. However, hs-CRP is more likely a risk marker rather than a causal factor of atherosclerosis [56,57]. The current data show that atorvastatin acts on the immune-response, offering a causal explanation on why statins ameliorate prognosis in ACS.

CONCLUSIONS
In ACS, ex-vivo atorvastatin treatment decreases the expression of transcription factors in T-cells, in particular the "immediate-early response protein" EGR1, resulting in inhibition of pro-inflammatory effector T-cells and activation of anti-inflammatory T-cells. EGR1 reduction by atorvastatin has also been confirmed in-vivo, after a single high-dose treatment. Therefore, in the setting of ACS, the early outcome improvement of intensive statin treatment might, at least partially, be related to direct inhibition of the master regulator EGR1 and to consequent immune-suppressive effects.
Some of the pathways inhibited by atorvastatin in the present study have recently been proposed as new therapeutic targets for the prevention of cardiovascular

Figure 10: Schematic representation of the effects of atorvastatin on CD4+T-cells in patients with ACS. Atorvastatin
decreases toll like receptor (TLR)-4 gene expression, a pattern-recognition receptor stimulated by several pathogen-associated molecular patterns (PAMPs) and damage associated molecular patterns (DAMPs), including bacterial lipolysaccaride (LPS) and ox-LDL, that has been implicated in the initiation and progression of atherosclerosis. TLR-4 stimulation induces intracellular pathways converging on nuclear factor (NF)-kB and mitogen-activated protein kinases (MAPK), with subsequent release of pro-inflammatory cytokines and expression of co-stimulatory molecules. Atorvastatin also reduces the expression of the chemokine (C-C motif) receptor 2 (CCR2), which is the receptor for monocyte chemoattractant protein (MCP)-1 and is involved in CD4 + T-cell transendothelial migration and recruitment at the site of tissue damage and inflammation. CCR2 intracellular pathway also converges on MAPK pathway, resulting in the activation of ERK and JNK and, eventually, of nuclear transcription factors FBJ murine osteosarcoma viral oncogene homolog (FOS) and early growth response 1 (EGR1), implicated in the immune response. Indeed, we observed an important decrease of EGR1 gene and protein expression by a single high-dose of atorvastatin both ex-vivo and in-vivo, suggesting a specific direct effects of atorvastatin on EGR1. The final net effect is a reduction of pro-inflammatory cytokine secretion and of chemokine and chemokine receptor synthesis, and an increase of anti-inflammatory pathways. *Green arrows indicate the effects of atorvastatin according to molecular assays (reduced expression of the TLR4, CCR2, FOS and EGR1 genes and of EGR1 protein); † black arrows indicate the opposite effects of atorvastatin on IFN-γ (reduced intracellular expression by CD4+CD28nullT-cells) and IL-10 (increased intracellular expression by CD4+CD25highT-cells) as assessed by flow-cytometry.
diseases. This is the case for the MCP-1/CCR2 pathway [58], the TLR4 pathway [59], and for a newly identified microRNA that functions as a negative regulator on inflammatory cytokines TNF-α and IL-6 via targeting EGR1 in vivo [60].

Study population
We prospectively evaluated 334 consecutive patients admitted to our CCU from June 2011 to June 2013 with a diagnosis of NSTE-ACS. We enrolled 40 patients (Cohort 1) who had never received statin treatment and with circulating CD4 + CD28 null T-cell frequency >4%, as our group has previously shown that CD4 + CD28 null T-cell frequency >4% predicts recurrence of acute coronary events [2]. Ten additional patients (Cohort 2) were enrolled from November 2013 to May 2014 using the same inclusion criteria, to analyze the in-vivo effects of atorvastatin on EGR1 gene expression and protein levels.
The study was approved by the local ethics committee and appropriate consent was obtained from study patients.

Isolation of CD4 + T-cells and atorvastatin treatment
Peripheral blood mononuclear cells (PBMCs) were obtained from whole blood samples by standard gradient centrifugation over Ficoll-Hypaque (GE Healthcare Bio-Sciences, Piscataway, NJ). CD4 + T-cells were isolated by magnetic micro-beads (CD4+T-cell isolation kit MACS, Miltenyi Biotec, Auburn, CA) according to the manufacturer's instructions. Pure atorvastatin, kindly provided by Pfizer (New York, NY), was dissolved in 2% dimethilsulfoxide (DMSO) solution at final concentration of 50mM. CD4 + T-cells were incubated for 24 hours, under sterile conditions at 37°C in an atmosphere containing 5% CO2, at a density of 1 x 10 6 /ml in RPMI 1640 medium (Sigma, St. Louis, MO) supplemented with 10% fetal bovin serum (Invitrogen, Carlsbad, CA) with and without increasing doses of atorvastatin: 3-10-26 μg/ml. These doses were chosen on the basis of the observation that comparable plasmatic concentrations correspond to in vivo doses of 10-40-80 mg of atorvastatin [61].

T-cell analysis by flow cytometry
In Cohort 1 patients, phenotypic and functional characteristics of CD4 + T-cells treated/untreated with atorvastatin were assessed by flow-cytometry using fluorochrome-conjugated monoclonal anti-human antibody (mAb), including isotype controls. Flow- For AGTCAGCCGCATCTTCTT Rev GCCCAATACGACCAAATCC Confirmation and validation of candidate genes among those reduced or increased by atorvastatin was performed by real time quantitative polymerase chain reaction (RT-qPCR) after reverse transcription of single-patient RNA.
CD4 + CD25 high T-cell frequency was obtained using monoclonal anti-CD4-FITC-labeled and monoclonal anti-human CD25-APC-labeled, on the basis of high CD25-APC fluorescence in comparison with intermediate CD25-APC fluorescence in CD4 + CD25 + T-cells. After cell surface staining, cell fixation and permeabilization, cells were stained with the intracellular PE-conjugated anti-Foxp3 mAB (eBioscience, San Diego, CA). The percentage of CD4 + CD25 high T-cells and of CD4 + CD25 high Tcells expressing Foxp3 + T-cells was then calculated as percentage of CD4 + CD25 + T-cell population.
In order to evaluate T-cell function, we determined IFN-γ production by CD4 + CD28 null T-cells and IL-10 production by CD4 + CD25 high T-cells upon stimulation. In the assessment of IL-10 production, regulatory T-cells were identified as CD4 + CD25 high T-cells, in order to avoid a stressful cell-handling as required to perform permeabilization for intra-cytoplamasmatic and intranuclear protein assessment, potentially resulting in a less accurate measurement of IL-10. A significant correlation was observed between CD4 + CD25 high T-cells and CD4 + CD25 high Foxp3 + T-cells (R = 0.67; P < 0.001). Briefly, cytokines production by CD4 + T-cell subsets was assessed after 4-hours in vitro activation with 100 ng/ml phorbol-2-myristate-13-acetate (PMA) (Sigma, St. Louis, MO) and 1µg/ml ionomycin (Sigma, St. Louis, MO). Cells were incubated in polypropylene tubes at 37°C for a total of 4 hrs; during the last 2 hrs, 10 µg/mL Brefeldin A (Sigma, St. Louis, MO) was added to block extracellular secretion of cytokines. After cell surface staining, cell fixation was done with IC FIX Buffer (eBioscience, San Diego, CA) for 15 min at 4°C. Cell membranes were reversibly permeabilized with Permeabilization Buffer (eBioscience, San Diego, CA) and intracellular cytokines were labeled with mouse anti-human IFN-γ PE-conjugated (eBioscience, San Diego, CA) and mouse anti-human IL-10 Peridinin-Chlorophyll-Protein-Complex (PerCP)-conjugated (R&D Systems, Minneapolis, MN). In addition, IFN-γ and IL-10 intracellular expression was presented as MFI.

Cytokine measurements
In the same patients of Cohort 1, citrate whole-blood samples were collected from an antecubital vein at the time of patient enrollment. Immediately after sampling, aliquots of 1 mL of whole blood were incubated for 24 hours under sterile conditions at 37°C in an atmosphere containing 5% CO2 without and with increasing doses of atorvastatin: 3, 10, 26 μg/ml. Afterwards, plasma samples were obtained, stored at -80°C and analysed in a single bath at the end of the study. Plasma levels of IL-10 and IFN-γ were measured with high-sensitivity ELISA kits (human-IL-10, Aushon Biosystems, Billerica, MA; human-IFN-γ, Bender MedSystem, Vienna, Austria), according to the manufacturer's instructions. The linear range of detection was 0.78 to 200 pg/ml for IL-10 and 1.6 to 100 pg/ml for IFN-γ. All samples were measured in duplicate, and the intra-and inter-assay variability was < 10%.

RNA preparation and quantitative PCR array analysis
Patients in Cohort 1 underwent quantitative PCR array. CD4 + T-cells were incubated for 24 hours without and with 26 µg/ml of atorvastatin (see above). Total RNA was isolated by using RNeasy minikit (Qiagen, Valencia, CA) from 3x10 6 CD4 + T-cells per patient. RNA was quantified using a Picodrop spectrophotometer measuring the absorbance at 260/280 nm, and RNA integrity was confirmed with the Agilent Bioanalyzer (Agilent Technologies, Santa Clara CA). For each condition (untreated/treated with atorvastatin) a pool of RNA was constituted, starting with the same amount of RNA (500 ng) from each patient.
Array targets were prepared from 435 ng of total RNA from each pool (untreated/treated with atorvastatin) using SABiosciences Kit for genomic DNA removal and reverse transcription, and two focused panels of genes were analyze by quantitative PCR array (RT² Profiler™ PCR Array, SABiosciences, Frederick, MD, USA). We performed the Human Transcription Factors RT² Profiler™ PCR Array (Catalog #PAHS-075) and the Human Th1-Th2-Th3 RT² Profiler™ PCR Array (Catalog #PAHS-034) both profiling the expression of 84 genes, according to the manufacturer's instruction. Thermal cycling parameters were 95°C for 10 min, followed by 40 cycles of amplifications at 95°C for 15 s, 55°C for 30 s, 72°C for 30 s, and 72°C for 5 min as the final elongation step performed on IQ5 Icycler (BioRad). Relative levels of mRNA expression were normalized in all samples with the expression levels of housekeeping genes, and data analysis was done using the Web portal (http:// www.sabiosciences.com/pcr/arrayanalysis.php).
The relative expression of each gene was compared with the expression in the control group and calculated using the ΔΔCT method. Each reported value represented the mean decrease (or increase) of mRNA expression relative to the control levels. A P-value of ≤0.05 and a fold change in gene expression of >3 were taken as significant.

Quantitative real-time RT-PCR
Confirmation and validation of candidate genes was performed by real time quantitative polymerase chain reaction (RT-qPCR) after reverse transcription of RNA obtained from each single patient of Cohort 1 using the iScript TM cDNA Synthesis Kit (BioRad Laboratories, Hercules, CA). For each patient 250 ng of RNA from CD4 + T-cells untreated/treated with atorvastatin was used to synthesize cDNA in a final volume of 20 µl. 1 µl of cDNA was used as template for RT-qPCR in a 15 µl reaction mixture, including SsoAdvanced TM SYBR ® Green supermix (BioRad) and 400 nm of each primer. Oligonucleotide primers for RT-qPCR were designed with Beacon Design (Table 4). RT-qPCR was performed on triplicates samples using the IQ5 Icycler (BioRad). After initial denaturation step of 30 sec at 95°C, a twosteps cycling procedure (denaturation at 95°C for 5 sec, annealing and extension at 64°C for 30sec) was performed for 40 cycles and followed by melting curve at 95°C for 6 sec. Data are normalized to human β-2-microglobulin (β-2MG) and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA levels as an endogenous control and are expressed relative to control sample using the formula 2 -∆∆CT , where C T is the threshold cycle number.
Also, the in-vivo effect of atorvastatin on EGR1 gene expression was assessed by RT-qPCR.

Western-blot assays
We also explored whether atorvastatin treatment might influence protein expression levels of EGR1, both ex-vivo and in-vivo. Western-blot analysis was carried out in CD4 + T-cells (1x10 6 ) cultured for 24 hours with and without 26 μg/ml atorvastatin. Western blot was performed using whole-cell extracts (25 μg per lane). Protein were resolved on 7% SDS-PAGE gels and transferred to nitrocellulose membranes. Membranes were blocked with 5% non-fat milk in phosphate buffered saline and 1% Tween-20 (TBS) and labeled with primary antibodies (all from Santa Cruz Biotechnology, Dallas, Tx): mouse monoclonal IgG 1 anti-human β-actin (sc-130301) and rabbit polyclonal IgG anti-human EGR1 p82 (588, sc-110). To saving samples, materials, and time we stripped and re-probed a single membrane for multiple targets instead of running and blotting multiple gels. Membranes were stripped with a buffer consisting of β-mercaptoethanol, SDS, and Tris-HCl and re-probed with the specific antibodies. Then, specific bands were detected by a chemiluminescent system (ChemiDoc MP System, Biorad) using corresponding secondary antibodies conjugated with horseradish peroxidase.
Also, the in-vivo effect of atorvastatin on EGR1 protein expression was assessed by western-blot.

In-vivo effects of atorvastatin
In-vivo effects of atorvastatin on EGR1 gene expression and protein levels were analyzed in Cohort 2 patients treated with atorvastatin 80 mg/daily, at 24 hours and 48 hours of treatment, by RT-qPCR and by westernblotting, respectively (see above).
We also determined the in-vivo effects of atorvastatin on IFN-γ production by CD4+CD28nullT-cells and IL-10 production by CD4+CD25highT-cells upon stimulation by flow cytometry (see above).

Statistical analysis
No power calculation could be performed because of lack of previous studies in this setting. Thus, the enrollment of 40 patients in Cohort 1 and of 10 patients in Cohort 2 was arbitrary.
The percentage of T-cell subsets along with cytokine production by these T-cells, were not normally distributed; they were expressed as median and range and analyzed using the Friedman's two-way analysis of variance by ranks for multiple pairwise comparisons with Dunnet's correction. The remaining continuous variables, including gene and protein expression were normally distributed; they were expressed as mean±SD and were compared using 1-way ANOVA for repeated measures with the Bonferroni correction for multiple pairwise comparisons or using the paired t-test, as appropriate. Proportions were compared using the Chi square test.
Univariate logistic regression analysis was applied to individuate the variables associated with the effects of atorvastatin (a reduction of CD4 + CD28 null T-cells producing IFNγ and/or an increase of CD4 + CD25 high T-cells producing IL-10 higher than 50% after incubation with atorvastatin 26 μg/ml). The following clinical and laboratory variables were tested: age, sex, classical risk factors, previous history of acute coronary events, left ventricular ejection fraction, multi-vessel disease, troponin T levels, lipid profile (total-cholesterol, LDL-cholesterol, and HDLcholesterol). At univariate logistic regression analysis none of the variables considered, including lipid profile, demonstrated any association with effects of atorvastatin. Therefore, the multivariate analysis was not performed.