Cardiovascular Research

Cardiovascular Research 2007 76(1):81-90; doi:10.1016/j.cardiores.2007.06.003

This Article Abstract FREE Full Text (PDF) E-letters: Submit a response Alert me when this article is cited Alert me when E-letters are posted Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to My Personal Archive Download to citation manager Request Permissions Disclaimer Google Scholar Articles by Turner, N. A. Articles by Porter, K. E. Search for Related Content PubMed PubMed Citation Articles by Turner, N. A. Articles by Porter, K. E. Social Bookmarking          What's this?

Copyright © 2007, European Society of Cardiology

Mechanism of TNF-induced IL-1, IL-1β and IL-6 expression in human cardiac fibroblasts: Effects of statins and thiazolidinediones

Neil A. Turner^a,*, Romana S. Mughal^a, Philip Warburton^a, David J. O'Regan^b, Stephen G. Ball^a and Karen E. Porter^a

^aAcademic Unit of Cardiovascular Medicine, Leeds Institute of Genetics, Health and Therapeutics (LIGHT), University of Leeds, Leeds LS2 9JT, UK
^bDepartment of Cardiac Surgery, The Yorkshire Heart Centre, Leeds General Infirmary, Leeds LS1 3EX, UK

*Corresponding author. Tel.: +44 113 3435890; fax: +44 113 3434803. n.a.turner{at}leeds.ac.uk

Received 29 November 2006; revised 4 June 2007; accepted 6 June 2007

	Abstract

Top Abstract 1. Introduction 2. Methods 3. Results 4. Discussion Acknowledgments References

 
Objective In addition to direct effects on myocardial cell function, tumor necrosis factor (TNF) contributes to adverse cardiac remodeling by increasing production of other pro-inflammatory cytokines [e.g. interleukin (IL)-1 and IL-6]. Both statins and thiazolidinediones (TZDs) have beneficial effects on cardiac remodeling, possibly due to their anti-inflammatory properties. The present study examined the mechanisms by which TNF stimulates expression of pro-inflammatory cytokines in cultured human cardiac fibroblasts and determined the effects of statin or TZD treatment.

Methods Human cardiac fibroblasts were cultured from biopsies of right atrial appendages. Cytokine mRNA expression and secretion was measured using quantitative real-time RT-PCR and ELISA. Activation of signaling pathways was determined by immunoblotting with phospho-specific antibodies.

Results TNF (0.1–10 ng/ml) stimulated IL-6, IL-1 and IL-1β mRNA expression in cardiac fibroblasts in a concentration-dependent manner. Pharmacological inhibitors and receptor-neutralizing antibodies established that both TNF-induced IL-6 and IL-1β expression was mediated via the TNFRI receptor and p38 mitogen-activated protein kinase (MAPK), phosphoinositide 3-kinase (PI3K)/Akt and nuclear factor (NF)-B pathways. In contrast, TNF-induced IL-1 expression required both TNFRI and TNFRII subtypes and p38 MAPK and PI3K/Akt pathways, but was negatively regulated by the NF-B pathway. Neither statins (simvastatin, fluvastatin) nor TZDs (ciglitazone, rosiglitazone, troglitazone) had inhibitory effects on TNF-induced IL-6 secretion or IL-1/β mRNA expression; indeed, cytokine expression was increased in response to TZDs.

Conclusions Our data provide important insights into the regulation of pro-inflammatory cytokine expression in human cardiac fibroblasts and suggest that the myocardial anti-inflammatory effects of statins and TZDs are not due to inhibition of TNF-induced IL-1 or IL-6 expression by cardiac fibroblasts.

KEYWORDS Cytokines; Interleukins; Receptors; Signal transduction; Statins

	1. Introduction

Top Abstract 1. Introduction 2. Methods 3. Results 4. Discussion Acknowledgments References

 
The importance of pro-inflammatory mediators in the development of heart failure has, over recent years, become well recognised. Following myocardial infarction (MI), the acute inflammatory phase serves initially to drive tissue repair and adaptation to injury [1]. However, in the longer term, sustained expression of pro-inflammatory cytokines is detrimental and can lead to progressive heart failure [1]. The pro-inflammatory cytokines tumor necrosis factor (TNF), interleukin-1 (IL-1) and interleukin-6 (IL-6) play key roles in this myocardial inflammatory response. In the early stages post-MI, inflammatory cells that infiltrate the infarct zone are the predominant source of these cytokines, however in the longer-term pro-inflammatory cytokine expression occurs at sites remote from the initial injury and originates from the myocardial cells themselves [2,3]. Cardiac fibroblasts are the most prevalent cell type in the human heart, accounting for two-thirds of all myocardial cells [4]. Both cardiomyocytes and cardiac fibroblasts have the capacity to secrete TNF and IL-6, but fibroblasts appear to be the principal source of IL-1 [5]. Cardiomyocytes respond to many pro-inflammatory cytokines by undergoing apoptosis and/or hypertrophy, whilst cardiac fibroblasts adopt a myofibroblast phenotype and undergo increased proliferation and collagen turnover, thus contributing to the adverse post-MI remodeling that can ultimately lead to heart failure [4].

Clinical studies have demonstrated a strong association between plasma TNF levels and the progression of left ventricular (LV) remodeling and heart failure [6,7]. Moreover, in animal models TNF infusion or cardiac-specific overexpression of TNF results in LV dysfunction and heart failure [8,9]. TNF exerts its biological effects via activation of two distinct cell surface receptors, the low affinity 55 kDa TNFRI receptor and the high-affinity 75 kDa TNFRII receptor [10], both of which are expressed in adult human myocardium [11]. Moreover, we have recently reported their expression in cultured adult human cardiac fibroblasts [12]. Most of the known biological effects of TNF are mediated via TNFRI activation, whereas the role of TNFRII is less well established [10].

Statins (HMG-CoA reductase inhibitors) are commonly prescribed cholesterol-lowering drugs that also appear to exert beneficial pleiotropic effects on myocardial remodeling independently of their lipid-lowering properties [13,14]. A potential mechanism by which statins can act in this way is by reducing expression of pro-inflammatory cytokines, as has been observed in the plasma of heart failure patients undergoing statin therapy [15,16]. More recent evidence from a rat MI model [17] and a study on heart transplant patients [18] suggests that statins can also reduce local expression of pro-inflammatory cytokines in the myocardium.

Thiazolidinediones (TZDs) are insulin-sensitizing agents that are primarily used for the treatment of Type 2 diabetic patients. However, in addition to improving insulin resistance, TZDs also exhibit a range of pleiotropic effects on cardiovascular cell function, including modulation of cell proliferation, migration, matrix remodeling and cytokine secretion [19,20]. TZDs ameliorate cardiac hypertrophy in animal models [21,22] and at the cellular level they can reduce cardiomyocyte hypertrophy [21,23] and modulate type I collagen expression by cardiac fibroblasts [24].

The aims of the present study were firstly, to determine whether TNF could stimulate expression of other pro-inflammatory cytokines (IL-6, IL-1 and IL-1β) in cultured human cardiac fibroblasts. Secondly, to conduct a comprehensive investigation of the intracellular mechanisms underlying these effects. Finally, to establish whether the reported anti-inflammatory effects of statins and TZDs on the myocardium could be explained by their ability to reduce pro-inflammatory cytokine expression by cardiac fibroblasts.

	2. Methods

Top Abstract 1. Introduction 2. Methods 3. Results 4. Discussion Acknowledgments References

 
2.1 Reagents
All cell culture reagents were purchased from Invitrogen (Paisley, Scotland, UK), except fetal calf serum that was from Biosera Ltd (Ringmer, East Sussex, UK). Human recombinant TNF was obtained from Sigma (Poole, Dorset, UK). All real-time RT-PCR reagents and Taqman probes/primers were from Applied Biosystems (www.appliedbiosytems.com). All statins, TZDs and signaling inhibitors were from Calbiochem (Nottingham, UK), except LY294002 that was from Alexis Biochemicals (Nottingham, UK).

2.2 Cell culture
Biopsies of right atrial appendage were obtained from patients without LV systolic dysfunction undergoing elective coronary artery bypass surgery at the Leeds General Infirmary. Local ethical committee approval and informed patient consent were obtained and the investigation conformed to the principles outlined in the Declaration of Helsinki, 1997. Primary cultures of cardiac fibroblasts were harvested, characterized as myofibroblasts (-smooth muscle actin- and vimentin-positive) and cultured as we have described previously [25,26]. Experiments were performed on cells from passages 3–5. Cells were serum-starved by incubation in serum-free medium (SFM) for 48 h prior to performing experiments in low serum medium (0.4% FCS).

2.3 Quantitative RT-PCR
Cells were pre-treated with low serum medium containing appropriate supplements (neutralizing antibodies or pharmacological agents) before addition of TNF. Cellular RNA was extracted after the appropriate incubation time using the Aurum Total RNA Kit (BioRad, Hemel Hempstead, UK) and cDNA prepared using the Reverse Transcription System (Promega, Southampton, UK), both according to the manufacturer's instructions. Real-time PCR was performed on duplicate samples using the Applied Biosystems 7500 Real-Time PCR System with human IL-6 (Hs00174131_m1), IL-1 (Hs00174092_m1) or IL-1β (Hs00174097_m1) primers and Taqman probes. Data are normalized to human GAPDH mRNA levels (Hs99999905_m1 primers) as an endogenous control (unaffected by any treatments used) and are expressed relative to either untreated control or TNF-treated sample using the formula 2^–""C_T, in which C_T is the threshold cycle number.

2.4 Neutralizing antibody experiments
Serum-starved fibroblasts were pre-treated for 1 h with monoclonal neutralizing antibodies (R...D Systems, Abingdon, Oxfordshire, UK) for TNFRI (MAB625) or TNFRII (MAB726) at a final concentration of 10 µg/ml prior to addition of TNF and measurement of cell signaling or cytokine mRNA levels.

2.5 Western blotting
Serum-starved fibroblasts were exposed to SFM containing 10 ng/ml TNF for the appropriate time before preparing whole cell homogenates and immunoblotting as we have described previously [27]. Activation of signaling pathways was determined using phospho-specific antibodies (Cell Signaling Technology, Hitchin, Hertfordshire, UK) and equal sample loading was confirmed using appropriate expression antibodies (Cell Signaling Technology). Immunolabelled bands were visualized by SuperSignal West Pico chemiluminescence kit (Perbio, Cramlington, Northumberland, UK).

2.6 Cytokine ELISAs
Cells were pre-treated for 4 h with low serum medium containing simvastatin or TZDs before addition of TNF for 24 h. Conditioned media were collected, centrifuged to remove cellular debris and stored at –40 °C for subsequent analysis. ELISAs were performed according to the manufacturer's instructions (R...D Systems). Samples for IL-6 analysis were diluted 1:50–1:100 and those for IL-1 and IL-1β were not diluted. The sensitivity (minimum detectable concentration) of the ELISAs was 0.70, 1.0 and 1.0 pg/ml for IL-6, IL-1 and IL-1β respectively.

2.7 Statistical analysis
Results are expressed as mean±SEM with n representing the number of experiments on cells from different patients. Differences between normalized data were analyzed using paired ratio t-tests. Dose response data were compared as ratios using repeated measures one-way analysis of variance (ANOVA) and Newman–Keuls post hoc test. All statistical analyses were performed using GraphPad Prism software (www.graphpad.com). P<0.05 was considered statistically significant.

	3. Results

Top Abstract 1. Introduction 2. Methods 3. Results 4. Discussion Acknowledgments References

 
3.1 Effect of TNF on pro-inflammatory cytokine expression in human cardiac fibroblasts
To determine whether TNF could induce expression of other pro-inflammatory cytokines in human cardiac fibroblasts, we employed quantitative real-time RT-PCR to simultaneously measure mRNA levels for IL-6, IL-1 and IL-1β following treatment of cells from 5 different patients with 0.1–10 ng/ml human TNF for 6 h. Steady-state mRNA levels of all three cytokines were increased in a concentration-dependent manner following TNF treatment, with maximal stimulation (5-fold, 7-fold and 11-fold increases for IL-6, IL-1 and IL-1β respectively) observed in response to 10 ng/ml TNF (Fig. 1). All subsequent experiments were performed using 10 ng/ml TNF.

View larger version (13K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 1 Effect of TNF on pro-inflammatory cytokine mRNA expression. Human cardiac fibroblasts were treated with 0.1–10 ng/ml TNF in low serum medium (0.4% FCS) for 6 h before extracting RNA. Relative quantification of cytokine mRNA levels was performed by real-time RT-PCR using specific Taqman primer/probes for IL-6 (A), IL-1 (B) and IL-1β (C). Data are expressed relative to control unstimulated sample (n=5). ANOVA: IL-6, IL-1 and IL-1β=P<0.0001. Newman–Keuls post hoc test: ***P<0.001, **P<0.01, *P<0.05 for effect of TNF.

 
3.2 Signaling pathways mediating TNF-induced cytokine expression
We have recently reported that TNF stimulates the extracellular signal-regulated kinase (ERK-1/2), p38 mitogen-activated protein kinase (p38 MAPK), phosphoinositide 3-kinase (PI3K)/Akt and nuclear factor-B (NF-B) pathways, but not the c-Jun NH₂-terminal kinase (JNK) pathway, in human cardiac fibroblasts [28]. To determine which of these signaling pathways was required for expression of individual cytokines in response to TNF, we utilized specific pharmacological inhibitors of the ERK kinase MEK-1 (PD98059), p38 MAPK (SB203580), PI3K (LY294002) and the IB kinase IKK-2 (IMD-0354).

Firstly, we established the effects of appropriate concentrations of these inhibitors by measuring phosphorylation of relevant downstream substrates, namely ERK-1/2, HSP27, Akt and IB- respectively. PD98059 (30 µM) and SB203580 (10 µM) inhibited the ERK and p38 MAPK pathways respectively, without affecting other signaling pathways (Fig. 2A). The PI3K inhibitor LY294002 (10 µM) readily inhibited TNF-induced Akt phosphorylation, and also reduced TNF-induced ERK phosphorylation (Fig. 2A). As LY294002 does not directly inhibit ERK activity [29], these data indicate that TNF-induced ERK phosphorylation occurs via a PI3K-dependent mechanism in human cardiac fibroblasts. Finally, the IKK-2 inhibitor IMD-0354 (10 µM) successfully prevented TNF-induced IB- phosphorylation and degradation (Fig. 2A). The TNF-induced reduction in IB- expression was also reversed by MG-132, an inhibitor of the proteasome degradation pathway, confirming that loss of signal was due to proteolytic degradation of IB- (data not shown). Additionally, IMD-0354 inhibited basal and TNF-induced ERK phosphorylation and increased basal HSP27 phosphorylation to levels similar to those observed in TNF-stimulated cells (Fig. 2A). Furthermore, IMD-0354 decreased basal Akt phosphorylation (Fig. 2A).

View larger version (42K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 2 Effect of pharmacological inhibitors on TNF-induced signaling and cytokine expression. Cells were pre-treated with PD98059 (30 µM), SB203580 (10 µM), LY294002 (10 µM) or IMD-0354 (10 µM) in low serum medium for 1 h prior to treatment with TNF (10 ng/ml). (A) Western blotting. Activation of specific signal transduction pathways was determined by immunoblotting whole cell homogenates with phospho-specific and total expression antibodies following 5 or 30 min TNF treatment. Antibodies used were phospho/total ERK-1/2 (ERK pathway), phospho/total HSP27 (p38 MAPK pathway), phospho/total Akt (PI3K/Akt pathway) and phospho/total IB- (NF-B pathway). Blots are representative of n=5. (B) RT-PCR. Cytokine mRNA levels were determined 6 h after TNF treatment. Data are expressed relative to control TNF-treated sample (n=6). +++P<0.001 for effect of TNF versus low serum control. ***P<0.001, **P<0.01, *P<0.05, NS= not significant for effect of inhibitor versus TNF alone.

 
Inhibitors were then employed to determine the role of these signaling pathways in mediating TNF-induced cytokine expression. TNF-induced IL-6 mRNA expression was reduced 60–80% by inhibitors of the p38 MAPK, PI3K and NF-B pathways, but was not affected by ERK pathway inhibition (Fig. 2B). In contrast, the TNF-induced increase in IL-1 mRNA levels was reduced by more than 80% by p38 MAPK and PI3K inhibition, unaffected by ERK inhibition, and increased 1.7-fold by NF-B inhibition (Fig. 2B). Finally, TNF-induced IL-1β mRNA expression was unaffected by ERK inhibition, partially reduced (35%) by p38 MAPK inhibition, and greatly reduced (80%) by PI3K/Akt and NF-B inhibition (Fig. 2B). As IMD-0354 modulated several signaling pathways (Fig. 2A), we also investigated the effects of two additional NF-B pathway inhibitors on TNF-induced cytokine expression. Both wedelolactone (25 µM) and MG-132 (10 µM) inhibited TNF-induced IL-6 mRNA expression to levels comparable with those following IMD-0354 treatment (data not shown). In contrast, wedelolactone (but not MG-132) mimicked the effect of IMD-0354 by increasing IL-1 mRNA levels and MG-132 (but not wedelolactone) mimicked the effect of IMD-0354 by reducing IL-1β mRNA levels (data not shown).

3.3 TNF receptor subtypes mediating TNF-induced cytokine expression
We have previously reported that human cardiac fibroblasts express both the TNFRI and TNFRII subtypes of the TNF receptor [12]. We therefore investigated which receptor subtypes mediated the effects of TNF on signaling and pro-inflammatory cytokine expression in these cells.

A TNFRI-selective neutralizing antibody inhibited TNF-induced phosphorylation of ERK-1/2, p38 MAPK and Akt, and IB- phosphorylation and degradation, whereas a TNFRII-selective neutralizing antibody had no effect on any of the pathways studied (Fig. 3A). JNK phosphorylation was not observed following TNF treatment or exposure to either neutralizing antibody (data not shown).

View larger version (44K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 3 Effect of TNF receptor-neutralizing antibodies on TNF-induced signaling and cytokine expression. Cells were pre-treated with TNFRI or TNFRII neutralizing antibodies (10 µg/ml) in low serum medium for 1 h before addition of 10 ng/ml TNF. (A) Western blotting. Whole cell homogenates were prepared 5 min (I-B phosphorylation) or 30 min (all other blots) after TNF exposure and immunoblots probed with phospho-specific and total expression antibodies. Blots are representative of n=4. (B, C, D) RT-PCR. Cellular RNA was extracted 6 h after TNF exposure and mRNA levels determined by RT-PCR. Data are expressed relative to TNF-treated sample (n=5). +++P<0.001, ++P<0.01 for effect of TNF versus low serum control. ***P<0.001, **P <0.01, NS= not significant for effect of neutralizing antibody versus TNF alone.

 
The identity of the TNF receptor subtype coupled to pro-inflammatory cytokine expression was determined by measuring TNF-induced cytokine mRNA levels following blockade with the same selective neutralizing antibodies. TNF-induced expression of IL-6 mRNA (Fig. 3B) and IL-1β mRNA (Fig. 3D) was inhibited specifically by the TNFRI neutralizing antibody, but not by the TNFRII antibody. In contrast, TNF-induced IL-1 mRNA expression was markedly inhibited (80%) by the TNFRI antibody and also partially inhibited (35%) by the TNFRII antibody (Fig. 3C). Basal levels of cytokine mRNA expression (i.e. in the absence of TNF) were increased slightly (1.5-to 2-fold) by the TNFRI antibody, but were not affected by the TNFRII antibody (data not shown). In subsequent experiments on cells from three different patients, we determined that simultaneous treatment with both neutralizing antibodies had no additional effects over those observed with the TNFRI antibody alone. Thus, the relative mRNA levels of IL-6, IL-1 and IL-1β were 0.70±0.01, 0.38±0.13 and 0.39±0.12 respectively in the presence of TNFRI antibody, compared with 0.75±0.11, 0.33±0.08 and 0.31±0.16 in the presence of TNFRI antibody plus TNFRII antibody.

3.4 Effects of statins on TNF-induced cytokine mRNA and secretion
Using real-time RT-PCR we determined that simvastatin, a commonly prescribed statin, reduced TNF-induced IL-6 mRNA levels by 45% (Fig. 4A). Although mRNA levels provide a useful indicator of effects on gene transcription and mRNA stability, they do not necessarily equate to changes in protein secretion, which may be subject to further downstream regulation (e.g. protein translation or vesicle trafficking). We therefore employed an ELISA to determine the effects of statins on TNF-induced IL-6 secretion in conditioned media after 24 h. Unstimulated cardiac fibroblasts constitutively secreted IL-6 (2600 pg/ml), and this was further increased 2.7-fold by TNF treatment (Fig. 4B). Despite reducing IL-6 mRNA levels, simvastatin had no effect on TNF-induced IL-6 secretion (Fig. 4B). Similar results were also obtained with another lipophilic statin, fluvastatin (Fig. 4B).

View larger version (17K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 4 Effect of statins on TNF-induced pro-inflammatory cytokine expression in human cardiac fibroblasts. Cells were pre-treated for 4 h with 10 µM simvastatin or 1 µM fluvastatin in low serum medium before exposure to TNF for 16 h (mRNA expression) or 24 h (ELISA). (A, C, D) RNA was extracted and real-time RT-PCR performed to determine IL-6, IL-1 and IL-1β mRNA levels. Data are expressed relative to control TNF-treated sample (n=5). +++P<0.001, ++P<0.01 for effect of TNF versus low serum control. *P<0.05 for effect of simvastatin versus TNF alone. (B) Conditioned media were collected and IL-6 concentration determined by ELISA (n=6). +++P<0.001 for effect of TNF versus low serum control. NS= not significant for effect of statin versus TNF alone.

 
The effect of simvastatin on TNF-induced IL-1 and IL-1β expression was then investigated. Simvastatin treatment consistently yielded a 3–4 fold increase in both TNF-induced IL-1 and IL-1β mRNA levels (Fig. 4C, D). ELISA analysis revealed the mean levels of IL-1 and IL-1β secretion from unstimulated cells after 24 h were only 0.03 pg/ml and 0.29 pg/ml respectively, which rose to 0.34 pg/ml and 0.57 pg/ml in TNF-treated cells. However, these levels remained well below the manufacturer's stated minimum detectable concentration for the ELISA of 1 pg/ml, indicating that secretion of these cytokines is negligible under these conditions.

3.5 Effects of TZDs on TNF-induced cytokine mRNA and secretion
The modulatory effect of three different TZDs (ciglitazone, rosiglitazone and troglitazone) on TNF-induced cytokine expression was determined. A trend towards increased TNF-induced IL-6 mRNA expression was evident following treatment of cardiac fibroblasts with ciglitazone or troglitazone, but rosiglitazone was without effect (Fig. 5A). A similar pattern was observed when IL-6 secretion was measured by ELISA, with both ciglitazone and troglitazone (but not rosiglitazone) stimulating a significant increase in TNF-induced IL-6 secretion (Fig. 5B). All three TZDs had stimulatory effects on TNF-induced IL-1 and IL-1β mRNA levels, although not all of these reached statistical significance (Fig. 5C, D).

View larger version (22K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 5 Effect of TZDs on TNF-induced pro-inflammatory cytokine expression in human cardiac fibroblasts. Cells were pre-treated for 1 h with 10 µM ciglitazone (CIG), rosiglitazone (ROS) or troglitazone (TRO) in low serum medium before exposure to TNF for 16 h (mRNA expression) or 24 h (ELISA). (A, C, D) RNA was extracted and real-time RT-PCR performed to determine IL-6, IL-1 and IL-1β mRNA levels. Data are expressed relative to control TNF-treated sample (n=6). +++P<0.001, ++P<0.01 for the effect of TNF versus low serum control. *P<0.05, NS= not significant for effect of TZD versus TNF alone. (B) Conditioned media were collected and IL-6 concentration determined by ELISA (n=6).+++P<0.001 for effect of TNF versus low serum control. ***P<0.001, **P<0.01, NS= not significant for effect of TZD versus TNF alone.

 

	4. Discussion

Top Abstract 1. Introduction 2. Methods 3. Results 4. Discussion Acknowledgments References

 
Our study reports several major findings. Firstly, that TNF increases mRNA expression of other pro-inflammatory cytokines (IL-6, IL-1 and IL-1β) in human cardiac fibroblasts in vitro. Secondly, that these effects are mediated via specific TNF receptors and signaling pathways (summarized in Fig. 6). Thirdly, that TNF induces a significant increase in IL-6 secretion, but not IL-1 or IL-1β secretion, from human cardiac fibroblasts. Finally, that TNF-induced IL-6 secretion is increased by the TZDs ciglitazone and troglitazone, but not modulated by statins or rosiglitazone.

View larger version (23K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 6 Summary of signaling networks coupling TNF to pro-inflammatory cytokine expression in human cardiac fibroblasts. Solid arrows indicate stimulatory role, thicker arrows represent greater role, and hatched lines indicate inhibitory role. TNF induces IL-6 expression via TNFRI, and is equally dependent on p38 MAPK, PI3K/Akt and NF-B activation, but independent of ERK activation. TNF-induced IL-1 expression occurs predominantly via TNFRI and is p38 MAPK- and PI3K/Akt-dependent, and is negatively regulated by NF-B. TNFRII is also involved in TNF-induced IL-1 expression, although the precise mechanism has yet to be established. TNF-induced IL-1β expression occurs solely via TNFRI and is fully dependent on PI3K/Akt and NF-B activation, partially dependent on p38 MAPK, but independent of ERK.

 
We have previously reported that human cardiac fibroblasts express both TNFRI and TNFRII receptor subtypes [12]. Although the use of soluble TNF favors activation of TNFRI compared with TNFRII [10], we revealed a novel role for TNFRII in mediating TNF-induced IL-1 expression. TNFRII was not coupled to activation of ERK, p38 MAPK, JNK, PI3K/Akt or NF-B pathways, as these were unaffected by the TNFRII neutralizing antibody. The modulation of IL-1 expression by TNFRII may be mediated via alternative TNFRII-specific signaling pathways, such as Bmx/Etk or STAT3 [30,31]. However, the lack of additive effect of the two neutralizing antibodies on TNF-induced IL-1 expression suggests a degree of interaction between the two receptor subtypes, possibly at the level of "ligand-passing" whereby the high-affinity TNFRII receptor binds TNF to present it to TNFRI [10]. That the TNFRI antibody only partially inhibited TNF-induced IL-6 mRNA expression (42% reduction) compared with IL-1/β (75–82% reduction) is likely due to incomplete inhibition of TNFRI, coupled with the differing sensitivities of expression of individual cytokines to TNF. Indeed, when the TNF concentration was reduced from 10 ng/ml to 1 ng/ml, comparable reductions in IL-6 (47%) and IL-1/β (81–84%) mRNA levels were observed (Fig. 1).

IL-6 gene transcription can be induced via several different intracellular signaling pathways, including ERK, p38 MAPK and NF-B [32]. In neonatal mouse cardiac fibroblasts, IL-6 expression was induced via activation of p38 MAPK, but not ERK, following stimulation of the β₂-adrenergic receptor or AMPK [33,34]. In contrast, angiotensin II-induced IL-6 production in neonatal rat cardiac fibroblasts required both p38 MAPK and ERK activation, but not NF-B [35], while G13 G-protein-induced IL-6 expression required NF-B activation [36]. In our present study, using human cardiac fibroblasts, TNF stimulated IL-6 mRNA expression via p38 MAPK, PI3K/Akt and NF-B pathways, but ERK activation was not important. It is clear therefore that induction of IL-6 expression in cardiac fibroblasts can proceed via multiple pathways, depending on the initiating stimulus.

Although IL-1 and IL-1β exert their functional responses via the same receptor, they are products of different genes, and their synthesis and regulation are very different [37]. The IL-1β gene promoter contains numerous transcription factor binding sites, including those for NF-B, AP-1, Sp1, NF-IL6 and CREB [37]. In the present study, TNF stimulated IL-1β mRNA expression via the p38 MAPK, PI3K/Akt and NF-B pathways, but not the ERK pathway. Sp1-mediated transcription has recently been shown to be stimulated by both the p38 MAPK and PI3K/Akt pathways [38,39], which may explain the importance of these pathways in promoting IL-1β transcription in our study. In contrast to IL-1β, relatively little is known about the regulation of IL-1 gene transcription. The IL-1 promoter lacks classical TATA and CAAT boxes and contains few consensus transcription factor binding sites, although sites for AP-1, Sp1 and a transcriptional repressor have been identified [40]. In the present study, inhibition of either the p38 MAPK or PI3K/Akt pathway prevented TNF-induced IL-1 expression, possibly also via inhibition of Sp1-mediated transcription. Interestingly, we also observed an increase in TNF-induced IL-1 expression following treatment with IMD-0354, an inhibitor of IB kinase-2 (IKK-2). Our data therefore propose a novel role for NF-B as a negative regulator of IL-1 gene expression in human cardiac fibroblasts.

Post-transcriptional regulation of IL-1 and IL-1β is also very different. IL-1 is not secreted from human cells, but accumulates intracellularly and at the plasma membrane [37]. A previous study on cultured human cardiac fibroblasts determined that IL-1 expression was localized to the nucleus and low level secretion occurred only after prolonged culture [41]. Our results concur with this and further demonstrate that despite observing 5- to 10-fold increases in steady state IL-1 mRNA levels, TNF treatment does not induce significant IL-1 secretion. Previous studies on monocytic cells revealed that IL-1β synthesis can also be regulated at the translational level [37]. Thus, several stimuli can induce significant amounts of IL-1β mRNA, without appreciable translation or secretion of IL-1β protein. This is likely the case in our study, in which only trace amounts of IL-1β were detected in culture medium after TNF treatment, despite >10-fold induction of IL-1β mRNA expression. This mechanism may be specific to human cardiac fibroblasts, as constitutive IL-1β secretion has been observed in unstimulated rat and mouse cardiac fibroblast cultures [36,42], whereas in cultured human cardiac fibroblasts it is absent (our data and [43]).

The ability of statins to reduce post-MI remodeling is likely a cumulative result of effects on cardiomyocytes and cardiac fibroblasts, the two major cell types of the heart. Statin treatment can reduce plasma levels of pro-inflammatory cytokines in heart failure patients [15,16] and reduce myocardial expression of pro-inflammatory cytokines in human hearts [18] and in animal models of MI [17]. However, the cellular source of cytokine expression affected by statin treatment in those studies was not determined. Our current findings that statin treatment does not reduce TNF-induced IL-6 secretion, and actually increases IL-1/β mRNA in human cardiac fibroblasts, suggest that the ability of statins to reduce myocardial cytokine expression is most likely due to effects on other cell types (e.g. cardiomyocytes).

TZDs, in addition to their insulin-sensitizing properties, possess anti-inflammatory properties and exhibit beneficial pleiotropic effects on both cardiomyocytes and cardiac fibroblasts that can act to reduce myocardial remodeling [20–24]. However, whether the ability of TZDs to reduce adverse myocardial remodeling can be explained by reduced myocardial pro-inflammatory cytokine expression has not previously been explored. Our data revealed that rather than reducing cytokine expression, two of the TZDs (ciglitazone and troglitazone) increased TNF-induced IL-6 mRNA expression and protein secretion, whereas rosiglitazone was without effect. Moreover, all three TZDs increased TNF-induced IL-1 and IL-1β mRNA expression. These potentially significant observations suggest that rather than being anti-inflammatory, TZDs may exert pro-inflammatory effects on cardiac fibroblasts that could exacerbate adverse myocardial remodeling. Although there is compelling evidence for beneficial myocardial effects of several TZDs in rodent MI models, this is not necessarily the case in larger animals. For example, rosiglitazone conferred no benefit and did not reduce myocardial IL-1 and IL-6 expression in a porcine ischemia/reperfusion model, although troglitazone was effective due to its pleiotropic anti-oxidant properties [44]. These data, together with our present findings, raise important questions regarding the choice of species as models for studying the effects of TZDs on the heart.

An unavoidable limitation of our study was its in vitro nature, and the exposure of cells to an artificial environment that differs from that in vivo. Nevertheless, our results provide important insights into how TNF, itself a potent pro-inflammatory cytokine, influences expression of other inflammatory stimuli in human cardiac fibroblasts, and how statins and TZDs may modulate this. A particular strength of our study was the use of cardiac fibroblasts from many different patients. Although precise patient characteristics were not available to us, the donors varied in age, gender, cardiovascular disease and drug treatment. Therefore our findings are representative of the bona fide human scenario. The minimum TNF concentration that induced significant IL-6 and IL-1 expression was 1 ng/ml, which is 20–30 times greater than that observed in the plasma of post-MI patients [45]. Given that local tissue levels may far exceed plasma levels, our findings are pertinent to the pathophysiological concentrations of TNF that occur in the human myocardium following MI. Fibroblasts were cultured from right atrial appendage; an area of the human heart that expresses high levels of TNF [46]. Whether ventricular fibroblasts would respond similarly to TNF is an issue requiring further study.

In summary, TNF acting via distinct TNF receptor subtypes and intracellular signaling pathways can stimulate the expression of IL-6, IL-1 and IL-1β by human cardiac fibroblasts. The inability of statins or TZDs to attenuate this inflammatory cascade indicates that cardiac fibroblasts are not the cellular targets for the anti-inflammatory effects of these drugs on the heart; indeed TZDs appear to potentiate some of the effects of TNF on pro-inflammatory cytokine expression. Our study reveals important new insights into how TNF, a cytokine strongly associated with heart failure progression, can lead to increased expression of other pro-inflammatory cytokines by human cardiac fibroblasts. A clearer understanding of the mechanisms of such effects may reveal novel targets for future therapeutic intervention.

Time for primary review 15 days

	Acknowledgments

Top Abstract 1. Introduction 2. Methods 3. Results 4. Discussion Acknowledgments References

 
This work was supported by the British Heart Foundation. NAT is in receipt of a Research Councils UK Academic Fellowship. We are grateful to Jean Kaye and Stacey Galloway for cell culture expertise.

	References

Top Abstract 1. Introduction 2. Methods 3. Results 4. Discussion Acknowledgments References

 

1. Nian M., Lee P., Khaper N., Liu P. Inflammatory cytokines and postmyocardial infarction remodeling. Circ Res (2004) 94:1543–1553.[Abstract/Free Full Text]
2. Irwin M.W., Mak S., Mann D.L., Qu R., Penninger J.M., Yan A., et al. Tissue expression and immunolocalization of tumor necrosis factor- in postinfarction dysfunctional myocardium. Circulation (1999) 99:1492–1498.[Abstract/Free Full Text]
3. Deten A., Volz H.C., Briest W., Zimmer H.G. Cardiac cytokine expression is upregulated in the acute phase after myocardial infarction. Experimental studies in rats. Cardiovasc Res (2002) 55:329–340.[Abstract/Free Full Text]
4. Jugdutt B.I. Ventricular remodeling after infarction and the extracellular collagen matrix: when is enough enough? Circulation (2003) 108:1395–1403.[Free Full Text]
5. Long C.S. The role of interleukin-1 in the failing heart. Heart Fail Rev (2001) 6:81–94.[CrossRef][Medline]
6. Deswal A., Petersen N.J., Feldman A.M., Young J.B., White B.G., Mann D.L. Cytokines and cytokine receptors in advanced heart failure: an analysis of the cytokine database from the vesnarinone trial (VEST). Circulation (2001) 103:2055–2059.[Abstract/Free Full Text]
7. Torre-Amione G., Kapadia S., Benedict C., Oral H., Young J.B., Mann D.L. Proinflammatory cytokine levels in patients with depressed left ventricular ejection fraction: a report from the Studies of Left Ventricular Dysfunction (SOLVD). J Am Coll Cardiol (1996) 27:1201–1206.[Abstract]
8. Bozkurt B., Kribbs S.B., Clubb F.J. Jr., Michael L.H., Didenko V.V., Hornsby P.J., et al. Pathophysiologically relevant concentrations of tumor necrosis factor- promote progressive left ventricular dysfunction and remodeling in rats. Circulation (1998) 97:1382–1391.[Abstract/Free Full Text]
9. Bryant D., Becker L., Richardson J., Shelton J., Franco F., Peshock R., et al. Cardiac failure in transgenic mice with myocardial expression of tumor necrosis factor-. Circulation (1998) 97:1375–1381.[Abstract/Free Full Text]
10. MacEwan D.J. TNF receptor subtype signalling: differences and cellular consequences. Cell Signal (2002) 14:477–492.[CrossRef][Web of Science][Medline]
11. Torre-Amione G., Kapadia S., Lee J., Bies R.D., Lebovitz R., Mann D.L. Expression and functional significance of tumor necrosis factor receptors in human myocardium. Circulation (1995) 92:1487–1493.[Abstract/Free Full Text]
12. Porter K.E., Turner N.A., O'Regan D.J., Ball S.G. Tumor necrosis factor induces human atrial myofibroblast proliferation invasion and MMP-9 secretion: inhibition by simvastatin. Cardiovasc Res (2004) 64:507–515.[Abstract/Free Full Text]
13. Reddy R., Chahoud G., Mehta J.L. Modulation of cardiovascular remodeling with statins: fact or fiction? Curr Vasc Pharmacol (2005) 3:69–79.[CrossRef][Medline]
14. van der Harst P., Voors A.A., van Gilst W.H., Bohm M., van Veldhuisen D.J. Statins in the treatment of chronic heart failure: biological and clinical considerations. Cardiovasc Res (2006) 71:443–454.[Abstract/Free Full Text]
15. Tousoulis D., Antoniades C., Vassiliadou C., Toutouza M., Pitsavos C., Tentolouris C., et al. Effects of combined administration of low dose atorvastatin and vitamin E on inflammatory markers and endothelial function in patients with heart failure. Eur J Heart Fail (2005) 7:1126–1132.[CrossRef][Web of Science][Medline]
16. Sola S., Mir M.Q., Lerakis S., Tandon N., Khan B.V. Atorvastatin improves left ventricular systolic function and serum markers of inflammation in nonischemic heart failure. J Am Coll Cardiol (2006) 47:332–337.[Abstract/Free Full Text]
17. Zhang J., Cheng X., Liao Y.H., Lu B., Yang Y., Li B., et al. Simvastatin regulates myocardial cytokine expression and improves ventricular remodeling in rats after acute myocardial infarction. Cardiovasc Drugs Ther (2005) 19:13–21.[CrossRef][Web of Science][Medline]
18. Wallace C.K., Stetson S.J., Kucuker S.A., Becker K.A., Farmer J.A., McRee S.C., et al. Simvastatin decreases myocardial tumor necrosis factor content in heart transplant recipients. J Heart Lung Transplant (2005) 24:46–51.[CrossRef][Web of Science][Medline]
19. Giannini S., Serio M., Galli A. Pleiotropic effects of thiazolidinediones: taking a look beyond antidiabetic activity. J Endocrinol Investig (2004) 27:982–991.[Web of Science][Medline]
20. Consoli A., Devangelio E. Thiazolidinediones and inflammation. Lupus (2005) 14:794–797.[Abstract/Free Full Text]
21. Asakawa M., Takano H., Nagai T., Uozumi H., Hasegawa H., Kubota N., et al. Peroxisome proliferator-activated receptor plays a critical role in inhibition of cardiac hypertrophy in vitro and in vivo. Circulation (2002) 105:1240–1246.[Abstract/Free Full Text]
22. Sakai S., Miyauchi T., Irukayama-Tomobe Y., Ogata T., Goto K., Yamaguchi I. Peroxisome proliferator-activated receptor- activators inhibit endothelin-1-related cardiac hypertrophy in rats. Clin Sci (Lond) (2002) 103(Suppl_48):16S–20S.[Medline]
23. Yamamoto K., Ohki R., Lee R.T., Ikeda U., Shimada K. Peroxisome proliferator-activated receptor activators inhibit cardiac hypertrophy in cardiac myocytes. Circulation (2001) 104:1670–1675.[Abstract/Free Full Text]
24. Chen K., Chen J., Li D., Zhang X., Mehta J.L. Angiotensin II regulation of collagen type I expression in cardiac fibroblasts: modulation by PPAR- ligand pioglitazone. Hypertension (2004) 44:655–661.[Abstract/Free Full Text]
25. Porter K.E., Turner N.A., O'Regan D.J., Balmforth A.J., Ball S.G. Simvastatin reduces human atrial myofibroblast proliferation independently of cholesterol lowering via inhibition of RhoA. Cardiovasc Res (2004) 61:745–755.[Abstract/Free Full Text]
26. Turner N.A., Porter K.E., Smith W.H., White H.L., Ball S.G., Balmforth A.J. Chronic β₂-adrenergic receptor stimulation increases proliferation of human cardiac fibroblasts via an autocrine mechanism. Cardiovasc Res (2003) 57:784–792.[Abstract/Free Full Text]
27. Turner N.A., Ball S.G., Balmforth A.J. The mechanism of angiotensin II-induced extracellular signal-regulated kinase-1/2 activation is independent of angiotensin AT_1A receptor internalisation. Cell Signal (2001) 13:269–277.[CrossRef][Web of Science][Medline]
28. Turner NA, Aley P, Hall KT, Warburton P, Galloway S, Midgley L, et al. Simvastatin inhibits TNF-induced invasion of human cardiac myofibroblasts via both MMP-9-dependent and-independent mechanisms J Mol Cell Cardiol in press, doi:10.1016/j.yjmcc.2007.05.006.
29. Vlahos C.J., Matter W.F., Hui K.Y., Brown R.F. A specific inhibitor of phosphatidylinositol 3-kinase 2-(4-morpholinyl)-8-phenyl-4H-1-benzopyran-4-one (LY294002). J Biol Chem (1994) 269:5241–5248.[Abstract/Free Full Text]
30. Luo D., Luo Y., He Y., Zhang H., Zhang R., Li X., et al. Differential functions of tumor necrosis factor receptor 1 and 2 signaling in ischemia-mediated arteriogenesis and angiogenesis. Am J Pathol (2006) 169:1886–1898.[Abstract/Free Full Text]
31. Theiss A.L., Simmons J.G., Jobin C., Lund P.K. Tumor necrosis factor (TNF) increases collagen accumulation and proliferation in intestinal myofibroblasts via TNF receptor 2. J Biol Chem (2005) 280:36099–36109.[Abstract/Free Full Text]
32. Vanden Berghe W., Vermeulen L., De Wilde G., De Bosscher K., Boone E., Haegeman G. Signal transduction by tumor necrosis factor and gene regulation of the inflammatory cytokine interleukin-6. Biochem Pharmacol (2000) 60:1185–1195.[CrossRef][Web of Science][Medline]
33. Du J.H., Xu N., Song Y., Xu M., Lu Z.Z., Han C., et al. AICAR stimulates IL-6 production via p38 MAPK in cardiac fibroblasts in adult mice: a possible role for AMPK. Biochem Biophys Res Commun (2005) 337:1139–1144.[Web of Science][Medline]
34. Yin F., Wang Y.Y., Du J.H., Li C., Lu Z.Z., Han C., et al. Noncanonical cAMP pathway and p38 MAPK mediate β₂-adrenergic receptor-induced IL-6 production in neonatal mouse cardiac fibroblasts. J Mol Cell Cardiol (2006) 40:384–393.[CrossRef][Web of Science][Medline]
35. Sano M., Fukuda K., Sato T., Kawaguchi H., Suematsu M., Matsuda S., et al. ERK and p38 MAPK but not NF-B are critically involved in reactive oxygen species-mediated induction of IL-6 by angiotensin II in cardiac fibroblasts. Circ Res (2001) 89:661–669.[Abstract/Free Full Text]
36. Nagamatsu Y., Nishida M., Onohara N., Fukutomi M., Maruyama Y., Kobayashi H., et al. Heterotrimeric G protein Ga13-induced induction of cytokine mRNAs through two distinct pathways in cardiac fibroblasts. J Pharmacol Sci (2006) 101:144–150.[CrossRef][Web of Science][Medline]
37. Dinarello C.A. Biologic basis for interleukin-1 in disease. Blood (1996) 87:2095–2147.[Abstract/Free Full Text]
38. D'Addario M., Arora P.D., McCulloch C.A. Role of p38 in stress activation of Sp1. Gene (2006) 379:51–61.[CrossRef][Web of Science][Medline]
39. Pore N., Liu S., Shu H.K., Li B., Haas-Kogan D., Stokoe D., et al. Sp1 is involved in Akt-mediated induction of VEGF expression through an HIF-1-independent mechanism. Mol Biol Cell (2004) 15:4841–4853.[Abstract/Free Full Text]
40. McDowell T.L., Symons J.A., Duff G.W. Human interleukin-1 gene expression is regulated by Sp1 and a transcriptional repressor. Cytokine (2005) 30:141–153.[CrossRef][Web of Science][Medline]
41. Heim A., Zeuke S., Weiss S., Ruschewski W., Grumbach I.M. Transient induction of cytokine production in human myocardial fibroblasts by coxsackievirus B3. Circ Res (2000) 86:753–759.[Abstract/Free Full Text]
42. Jaffre F., Callebert J., Sarre A., Etienne N., Nebigil C.G., Launay J.M., et al. Involvement of the serotonin 5-HT_2B receptor in cardiac hypertrophy linked to sympathetic stimulation: control of interleukin-6 interleukin-1β and tumor necrosis factor- cytokine production by ventricular fibroblasts. Circulation (2004) 110:969–974.[Abstract/Free Full Text]
43. Ancey C., Corbi P., Froger J., Delwail A., Wijdenes J., Gascan H., et al. Secretion of IL-6 IL-11 and LIF by human cardiomyocytes in primary culture. Cytokine (2002) 18:199–205.[CrossRef][Web of Science][Medline]
44. Xu Y., Gen M., Lu L., Fox J., Weiss S.O., Brown R.D., et al. PPAR- activation fails to provide myocardial protection in ischemia and reperfusion in pigs. Am J Physiol Heart Circ Physiol (2005) 288:H1314–H1323.[Abstract/Free Full Text]
45. Kosmala W., Przewlocka-Kosmala M., Mazurek W. Proinflammatory cytokines and myocardial viability in patients after acute myocardial infarction. Int J Cardiol (2005) 101:449–456.[CrossRef][Web of Science][Medline]
46. Valen G., Paulsson G., Bennet A.M., Hansson G.K., Vaage J. Gene expression of inflammatory mediators in different chambers of the human heart. Ann Thorac Surg (2000) 70:562–567.[Abstract/Free Full Text]

CiteULike    Connotea    Del.icio.us    What's this?

This article has been cited by other articles:

				X. Palomer, D. Alvarez-Guardia, R. Rodriguez-Calvo, T. Coll, J. C. Laguna, M. M. Davidson, T. O. Chan, A. M. Feldman, and M. Vazquez-Carrera TNF-{alpha} reduces PGC-1{alpha} expression through NF-{kappa}B and p38 MAPK leading to increased glucose oxidation in a human cardiac cell model Cardiovasc Res, March 1, 2009; 81(4): 703 - 712. [Abstract] [Full Text] [PDF]

				A. I. Totlandsdal, M. Refsnes, T. Skomedal, J.-B. Osnes, P. E. Schwarze, and M. Lag Particle-Induced Cytokine Responses in Cardiac Cell Cultures--the Effect of Particles versus Soluble Mediators Released by Particle-Exposed Lung Cells Toxicol. Sci., November 1, 2008; 106(1): 233 - 241. [Abstract] [Full Text] [PDF]

				Y. Park, S. Capobianco, X. Gao, J. R. Falck, K. C. Dellsperger, and C. Zhang Role of EDHF in type 2 diabetes-induced endothelial dysfunction Am J Physiol Heart Circ Physiol, November 1, 2008; 295(5): H1982 - H1988. [Abstract] [Full Text] [PDF]

				J. L. Guy, D. W. Lambert, A. J. Turner, and K. E. Porter Functional angiotensin-converting enzyme 2 is expressed in human cardiac myofibroblasts Exp Physiol, May 1, 2008; 93(5): 579 - 588. [Abstract] [Full Text] [PDF]

This Article Abstract FREE Full Text (PDF) E-letters: Submit a response Alert me when this article is cited Alert me when E-letters are posted Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to My Personal Archive Download to citation manager Request Permissions Disclaimer Google Scholar Articles by Turner, N. A. Articles by Porter, K. E. Search for Related Content PubMed PubMed Citation Articles by Turner, N. A. Articles by Porter, K. E. Social Bookmarking          What's this?

Online ISSN 1755-3245 - Print ISSN 0008-6363

Copyright © 2009 European Society of Cardiology

Oxford Journals Oxford University Press

• Site Map
• Privacy Policy
• Frequently Asked Questions

Other Oxford University Press sites: Oxford University Press Oxford Journals China Oxford Journals Japan American National Biography Booksellers' Information Service Children's Fiction and Poetry Children's Reference Corporate & Special Sales Dictionaries Dictionary of National Biography Digital Reference English Language Teaching Higher Education Textbooks Humanities International Education Unit Law Medicine Music Online Products Oxford English Dictionary Reference Rights and Permissions Science School Books Social Sciences Very Short Introductions World's Classics