A Soluble Factor from Trypanosoma cruzi Inhibits Transforming Growth Factor-ß-Induced MAP Kinase Activation and Gene Expression in Dermal Fibroblasts

The protozoan parasite Trypanosoma cruzi, which causes human Chagas' disease, exerts a variety of effects on host extracellular matrix (ECM) including proteolytic degradation of collagens and dampening of ECM gene expression. Exposure of primary human dermal fibroblasts to live infective T. cruzi trypomastigotes or their shed/secreted products results in a rapid down-regulation of the fibrogenic genes collagenIα1, fibronectin and connective tissue growth factor (CTGF/CCN2). Here we demonstrate the ability of a secreted/released T. cruzi factor to antagonize ctgf/ccn2 expression in dermal fibroblasts in response to TGF-ß, lysophosphatidic acid or serum, where agonist-induced phosphorylation of the mitogen-activated protein (MAP) kinases Erk1/2, p38 and JNK was also inhibited. Global analysis of gene expression in dermal fibroblasts identified a discrete subset of TGF-ß-inducible genes involved in cell proliferation, wound repair, and immune regulation that are inhibited by T. cruzi secreted/released factors, where the genes exhibiting the highest sensitivity to T. cruzi are known to be regulated by MAP kinase-activated transcription factors. Consistent with this observation, the Ets-family transcription factor binding site in the proximal promoter region of the ctgf/ccn2 gene (−91 bp to −84 bp) was shown to be required for T. cruzi-mediated down-regulation of ctgf/ccn2 reporter expression. The cumulative data suggest a model in which T. cruzi-derived molecules secreted/released early in the infective process dampen MAP kinase signaling and the activation of transcription factors that regulate expression of fibroblast genes involved in wound repair and tissue remodelling, including ctgf/ccn2. These findings have broader implications for local modulation of ECM synthesis/remodelling by T. cruzi during the early establishment of infection in the mammalian host and highlight the potential for pathogen-derived molecules to be exploited as tools to modulate the fibrogenic response.


Introduction
The kinetoplastid protozoan parasite Trypanosoma cruzi causes Chagas' disease in humans, a chronic and debilitating condition affecting several million individuals in Latin America. T. cruzi is transmitted by an insect vector which gains access to the host via breaches in the skin or through mucosal membranes, mainly conjunctival or gastric mucosa [1,2]. As an obligate intracellular parasite that disseminates from initial infection sites to tissues such as heart and smooth muscle, T. cruzi undergoes multiple rounds of invasion, growth and egress from infected cells during the acute stage of infection. Very little is currently known regarding the early interactions between T. cruzi and its host that facilitate establishment of infection in vivo. Cellular models of T. cruzi infection have been very useful for defining the molecular and cellular events that regulate the early parasite-host cell interactions and host cell invasion. During its early interaction with mammalian host cells, trypomastigotes, the invasive forms of T. cruzi, trigger rapid changes in a number of cellular signaling pathways to facilitate the process of parasite entry into non-professional phagocytic cells (reviewed in [3,4]). While these early signaling events have been relatively well-studied in the context of T. cruzi invasion, little is known regarding the impact of these parasite-induced signaling cascades downstream of the invasion process. Transcriptional profiling of T. cruzi-infected fibroblasts revealed that the earliest detectable changes triggered by infective T. cruzi trypomastigotes involve down-regulation of a small subset of genes including members of the CCN family (cyr61 and ctgf/ccn2) [5], which play important roles in angiogenesis and extracellular matrix (ECM) homeostasis [6]. T. cruzi-dependent dampening of ctgf/ccn2 expression occurs at both the mRNA and protein levels and is mediated by a secreted/released parasite factor that is capable of antagonizing TGF-ß-mediated induction of ctgf/ccn2 [7].
Connective tissue growth factor (CTGF/CCN2) is a 38 kDa secreted cysteine-rich heparin-binding glycoprotein [8] that promotes cell proliferation and co-operates with TGF-ß to promote myofibroblast differentiation and enhanced extracellular matrix (ECM) synthesis (reviewed in [9]). Dysregulation of CTGF/CCN2 expression leads to excessive scarring and fibrosis and this cytokine is over-expressed in a variety of tumors where CTGF/CCN2 levels correlate with disease progression [9]. As such there has been significant interest in CTGF/CCN2 as a therapeutic target for a number of disease states [10,11,12,13]. Our finding that the human pathogen, Trypanosoma cruzi, releases a factor that inhibits TGF-ß-mediated expression of CTGF/CCN2 prompted further investigation into the mechanistic basis for this observation. CTGF/CCN2 expression is induced by diverse extracellular stimuli, including growth factors, cytokines and mechanical stress [9,14,15,16,17]. TGF-ß-stimulated expression of CTGF/CCN2 requires the activation of SMAD proteins and MAP kinases downstream of the TGF-ß receptor [9,15]. Erk1/2 and p38 are generally associated with positive regulation of CTGF/CCN2 expression in different cell types [15,18,19,20], whereas the role of JNK is more variable [15,18,20]. The expression of CTGF/CCN2 is also controlled via the activities of the ETS family of transcription factors. A functional Ets-binding site identified in the proximal ctgf/ccn2 promoter, spanning the region 291 to 284 bp upstream of the transcriptional start site, is bound by Ets-1 to promote TGF-ß-dependent induction of ctgf/ ccn2 [21], while binding of the same site by Fli-1 negatively regulates ctgf/ccn2 expression in human fibroblasts [22]. The activity of many members of the ETS family is controlled through MAP kinase signaling, where the DNA-binding and transactivation activities of ETS transcription factors are regulated by phosphorylation [23,24].
In the present study, we demonstrate that T. cruzi-dependent abrogation of ctgf/ccn2 expression in human dermal fibroblasts is associated with inhibition of both basal and agonist-induced activation of MAP kinase signaling and requires the functional Etsbinding site in the proximal promoter of the ctgf/ccn2 gene. Expanding our analysis of the impact of T. cruzi released factors on TGF-ß-induced fibroblast gene expression we describe a discrete subset of agonist-inducible fibroblast genes that are sensitive to T. cruzi secreted/released factors. We report that the group of TGFß-inducible genes that exhibit the highest sensitivity to a T. cruzi secreted/released fraction are MAP kinase-regulated genes that function in wound repair, extracellular matrix remodelling and host response pathways. Collectively, these findings provide novel insights into early T. cruzi-host cell interactions. We demonstrate the ability of secreted/released parasite molecules to hinder agonist-induced host cell signaling pathways and gene expression. If similar events occur early in the T. cruzi infection process in vivo, it would be predicted that local inhibition of ECM synthesis would facilitate dissemination from initial sites of infection.

Results
A secreted/released trypanosome factor inhibits agonistinduced expression of CTGF/CCN2 in human dermal fibroblasts Connective tissue growth factor (CTGF/CCN2) expression is rapidly repressed in human dermal fibroblasts infected with the protozoan parasite, Trypanosoma cruzi, where the CTGF/CCN2repressive activity was shown to be associated with a secreted/ released trypanosome factor that is present in parasite-conditioned medium (PCM) [5] [7]. While the repressive factor released from live infective T. cruzi trypomastigotes into the medium (parasiteconditioned medium; PCM) [7] has not been identified, it is associated with a trypsin-sensitive, heat-labile, high molecular weight protein fraction of T. cruzi PCM enriched in GPI-anchored surface proteins (A. Mott and B. Burleigh, unpublished data). Here we show that treatment of human dermal fibroblasts (HFF) with T.
cruzi PCM blocks induction of ctgf/ccn2 expression triggered by different agonists TGF-ß ( Figure 1A, B), serum ( Figure 1C) or lysophosphatidic acid (LPA) ( Figure 1D). Thus, the ability of T. cruzi PCM to inhibit ctgf/ccn2 expression in response to distinct exogenous agonists suggests that signaling molecules primarily associated with TGF-ß receptor signaling such as the SMADs, are unlikely to be the downstream target of T. cruzi PCM leading to the repression of ctgf/ccn2.

T. cruzi PCM inhibits agonist-induced MAP kinase activation
Given the well-established role of MAP kinase signaling in the regulation of ctgf/ccn2 expression in a variety of cell types [15,18,19,20,25,26], Erk1/2, JNK and p38, were targeted in HFF with selective inhibitors to determine the impact on TGF-ß-inducible expression of ctgf/ccn2. Added individually, inhibitors of Erk1/2, JNK and p38 partially inhibited TGF-ß-dependent up-regulation of ctgf/ccn2 whereas in combination, the inhibitors were much more effective (Figure 2A) suggesting that different arms of the MAP kinase signaling pathway are involved in regulation of ctgf/ ccn2 expression in HFF. As predicted from this result, stimulation of HFF with TGF-ß (5 ng/ml) or LPA (10 mM) promotes increased phosphorylation of Erk1/2, JNK and p38, which migrate on western blots at the expected molecular weights of 40 kDa, 45 kDa and 39 kDa respectively, where T. cruzi PCM abrogates both TGF-ß and LPA-induced phosphorylation of these kinases ( Figure 2B). Combined, the data suggest that T. cruzi PCM exerts an inhibitory effect on ctgf/ccn2 expression by dampening MAP kinase signaling pathways.
An Ets binding site in the proximal ctgf/ccn2 promoter is involved in T. cruzi-mediated repression of gene expression The ctgf/ccn2 promoter/enhancer region contains several protein-binding sites that have been implicated in regulation of gene expression [21,22,27,28,29,30]. Focusing on the proximal promoter region of ctgf/ccn2 ( Figure 3A), we tested the impact of T. cruzi PCM treatment on basal and TGF-ß stimulated reporter activity using a series of ctgf/ccn2 promoter/SEAP reporter plasmids [21,27] transfected into human dermal fibroblasts. As previously described, reporters containing functional Smad3 and Ets-binding sites: ie. spanning regions -805 to +17 bp ( Figure 3B; 2805) or 2244 to +17 bp ( Figure 3B ; 2244) of the ctgf/ccn2 promoter were activated by TGF-ß whereas a severely truncated promoter/reporter spanning 286 to +17 bp of the ctgf/ccn2 promoter ( Figure 3B; 286) failed to activate following TGF-ß stimulation. T. cruzi PCM treatment resulted in a significant decrease in basal activity of the 2805 and 2244 reporter constructs and blocked TGF-ß mediated activation of these reporters ( Figure 3B). In contrast, basal expression of the reporter from the truncated construct ( Figure 3B; 286) did not change in response to T. cruzi PCM. These data suggest that the Ets site present at 291 to 284 bp upstream of the transcriptional start site in the promoter is required to mediate the repressive effects of T. cruzi PCM on ctgf/ccn2 gene expression. In order to confirm this critical role for the Ets consensus site in PCM-mediated repression of ctgf/ccn2 expression, we examined the effect of T. cruzi PCM on the full-length reporter containing a mutated Ets site ( Figure 3B; 2805 ETS). As previously reported, a functional Ets binding site is required for TGF-ß-stimulated ctgf/ccn2 gene expression, and therefore this reporter is refractory to TGF-ß treatment ( Figure 3B). Critically, a comparison between the response of the 2805 and the mutant 2805 ETS to PCM demonstrates the inability of PCM to repress activity in the absence of a functional Ets binding site ( Figure 3B). We conclude from these experiments that a functional Ets binding site in the ctgf/ccn2 promoter is essential for TGF-ß dependent gene expression, and imparts sensitivity to T. cruzimediated inhibition of ctgf/ccn2 gene expression.

T. cruzi PCM inhibits the expression of a subset of MAP kinase-regulated fibroblast genes involved in wound repair
Results presented above demonstrate that the factor(s) that are secreted or shed by T. cruzi trypomastigotes antagonize MAP kinase signaling in HFF downstream of TGF-ß and LPA. To determine the global impact of T. cruzi PCM on both basal and TGF-ß stimulated gene expression in human dermal fibroblasts, transcriptomic analyses were conducted following exposure of HFF to TGF-ß (5 ng/ml) in the presence or absence of T. cruzi PCM for 3 hours. cDNA prepared from treated and control cells were hybridized to human Affymetrix arrays containing ,47,000 probesets (HG_U_133, 2.0). Raw expression data from triplicate experiments was imported into, and processed using, Rosetta Resolver 7.0 as outlined in the Experimental Procedures. Genes with expression levels that differed by 2-fold or greater (p,0.01) between the different experimental groups were considered for further analysis.
Examining first the impact of the secreted/released T. cruzi fraction on fibroblast gene expression, we find 89 unique genes (in 121 probe-sets) consistently down-regulated in PCM-treated cells and 82 genes (in 107 probe-sets) up-regulated greater than 2-fold ( Table S1). Many of the repressed genes are 'immediate early' genes that regulate transcription (egr-1, jun, myc, fos, KLF2, KLF6) and are involved in cell proliferation [31,32,33,34,35] and inflammation [36,37,38]. As predicted from our studies with ctgf/ccn2, T. cruzi PCM treatment also results in the down-regulation of genes that are involved in extracellular matrix synthesis and tissue remodeling including hyaluronan synthase 2 (HAS2), tenascin C (TNC) and the CCN family members ctgf/ccn2 and cyr61 [17].
Biological network analysis conducted using the Ingenuity Pathway Analysis TM software program reveals a dense set of relationships between the subset of T. cruzi PCM sensitive T. cruzi PCM abrogates MAP kinase activation and decreased MAP kinase signaling results in inhibition of ctgf/ccn2 expression. (A) HFF were stimulated with 5 ng/ml TGF-ß1 in the presence of medium, T. cruzi PCM or with 10 mM SP600125, 20 mM SB203580 or 50 mM PD98059 as indicated for 2 hours prior to mRNA harvest for quantitative real-time PCR analysis. MAP kinase inhibitor treatments were carried out following a 30-minute pre-incubation step. Data is represented as the mean 6 s.e. from triplicate experiments (n = 4). Statistical significance was assessed using the Student's t-test (** p,0.01, * p,0.05). (B) Western blot of phospho-Erk, phospho-JNK and phospho-p38 normalized to total Erk, JNK and p38 respectively in lysates from HFF stimulated with 5 ng/ml TGF-ß1 or 10 mM LPA for 5 or 15 minutes in serum-free media or parasiteconditioned medium. Results for densitometric analysis are shown as numerical values below each panel and represented as phosphorylation relative to mock-treated controls (arbitrarily set to a value of 1.0). doi:10.1371/journal.pone.0023482.g002 TGF-ß-inducible genes and MAP kinases ( Figure 5) where the top functions associated with this network include 'cellular growth', 'cancer' and 'connective tissue development and function'. A comparative analysis of the top canonical pathways represented in the T. cruzi PCM-sensitive and PCM-insensitive TGF-ß inducible gene datasets are shown in Table S3. The top canonical pathways associated with the PCM-sensitive genes reveal 5 pathways with significant p-values (p,0.05) including 'hepatic fibrosis/hepatic stellate cell activation' (CTGF, FN1, TGFb2, EDN1, CCL2, IL-6) as the most significant pathway affected by the secreted/shed T. cruzi fraction. In contrast, when the PCM-insensitive group of TGF-ßinducible genes was considered, the top functions involved 'axonal guidance' and 'clathrin-mediated endocytosis signaling' (Table S4) [43,44,45,46,47,48,49]. Collectively, the global expression data are supportive of a model in which infective T. cruzi trypomastigotes secrete or release factors that within minutes dampen MAP kinase signaling in human dermal fibroblasts. We propose that this dampening of host cell MAP kinase activity negatively impacts the activities of downstream transcription factors such as the ETS family of transcription factors known to be involved in the regulation of a number of genes involved in ECM synthesis and wound repair [50,51] including ctgf/ccn2 gene expression [21,22,51].

Discussion
This study demonstrates that mammalian-infective forms of Trypanosoma cruzi release a factor (or factors) that significantly impacts host cell signaling cascades, thereby altering the expression of a subset of genes involved in cell proliferation, wound repair and inflammation. Focusing on the pro-fibrogenic cytokine, CTGF/CCN2, a multifunctional secreted protein with a central role in wound repair and fibrosis [9], we demonstrate that a shed/secreted T. cruzi factor inhibits both basal and agonistinduced up-regulation of ctgf/ccn2 gene expression. In line with previous reports demonstrating a critical role for MAP kinases in regulating ctgf/ccn2 gene expression in response to pro-fibrogenic agonists [15,17,18,19,20], our data demonstrate a role for Erk1/2, JNK and p38 in TGF-ß-mediated ctgf/ccn2 expression in low passage primary HFF. Exposure of HFF to the shed/secreted T. cruzi fraction inhibits basal and agonist-dependent phosphorylation of Erk1/2, JNK and p38 and the induction of ctgf/ccn2 expression, a downstream target of MAP kinase activation [15]. Analysis of the global transcriptional response to T. cruzi PCM in HFF revealed significant repression of a subset of TGF-ß-inducible genes involved in cell proliferation, wound healing and inflammatory responses, where the genes most sensitive to T. cruzi-mediated inhibition (EGR1, HBEGF, PTGS2, LIF, EDN1, IL6, and CYR61) are known to be regulated by MAP kinase-activated transcription factors such as the ETS-family members and AP-1 [43,44,45,46,47,48,49]. In addition, genes encoding components of the AP-1 complex, c-jun and c-fos, as well as egr-1 were also rapidly down-regulated in T. cruzi PCM treated cells, suggesting the disruption of an important regulatory network centered around MAP kinase signaling, which is required for efficient expression of ctgf/ccn2 and a host of other TGF-ß-inducible genes. Consistent with the notion that T. cruzi might negatively regulate MAP kinase-activated transcription factors, we demonstrated that a functional Ets-binding site (spanning 291 and 284 bp) in the upstream ctgf/ccn2 promoter region, which is critical for TGF-ß stimulated expression of ctgf/ccn2 [15,21,22], is required for T. cruzi PCM repression of both basal and TGF-ß-stimulated reporter expression. Thus, we propose that the observed inhibition of agonist-induced phosphorylation of the MAP kinases, Erk, p38 and JNK, would impact the activation of MAP kinase-activated transcription factors such as Ets-1. Failure to assemble Ets transcription factor(s) at the Ets binding site, possibly in conjunction with other accessory proteins such as AP-1 [21], would prevent activation of the ctgf/ccn2 promoter in response to signals mediated by exogenous agonists such as TGF-ß [15]. Overall, our findings reinforce recent literature regarding the critical role for MAP kinase signaling in regulation of ctgf/ccn2 expression [9,18,20] as well as for the role of the Ets-binding site on the ctgf/ccn2 promoter [21,22] in driving ctgf/ccn2 gene expression. Given that CTGF/CCN2 is a therapeutic target for the prevention and treatment of a number of disease states in which the expression of this critical cytokine is dysregulated [10,11,12,13], our finding that the mammalian-infective forms of T. cruzi release a factor that interferes with agonist-dependent up-regulation of CTGF/CCN2 warrants further study.
We had shown previously that infection with live T. cruzi trypomastigotes or parasite-conditioned medium (PCM) causes rapid dephosphorylation of Erk1/2 and sustained repression of extracellular matrix protein expression in dermal fibroblasts [7]. Focusing on the secreted/released T. cruzi fraction in this study, we demonstrate the broad impact of T. cruzi PCM on MAP kinase signaling where agonist-dependent activation of Erk1/2, p38 and JNK is thwarted, as is the expression of a subset of TGF-ßdependent gene expression. Our findings are consistent with a   [52]. While these in vitro observations provide novel insights into the impact of infective T. cruzi trypomastigotes on host cell signaling and gene expression, the ultimate goal would be to measure these events and outcomes in the context of an in vivo infection. Given the critical role played by MAP kinases in cell proliferation, wound healing and host immunity [19,53,54,55]. It is tempting to speculate that modulation of this signaling pathway early in the T. cruzi infection process could modulate the local host environment in favor of the parasite. If, as suggested by in vitro studies, ECM synthesis and host immune response pathways were to be repressed at local sites of T. cruzi infection, we would predict that these events would facilitate the early establishment of infection in the host. While this would be challenging to test in the context of ECM homeostasis, the concept of immune evasion early in the establishment of infection by T. cruzi trypomastigotes is not new [56]. There are several examples of soluble T. cruzi immunomodulatory molecules including the trypanosomal immunosuppressive factor (TIF) which represses T-and B-cell functions [57,58,59,60] and a mucin-like protein expressed on the surface of epimastigotes and metacyclic trypomastigotes (AgC10) which inhibits LPS-mediated signaling and dendritic cell activation via MAP kinase [61]. A parasite glycoinositol-phospholipid [62] and another polypeptide antigen (Ag123) [63] have also been implicated in modulating the host immune response demonstrating that T. cruzi may have evolved several overlapping mechanisms for interfering with normal host immune function. The T. cruzi molecule(s) responsible for antagonizing TGF-ß signaling in dermal fibroblasts, that dampen fibrogenic and immune gene expression, have yet to be identified. Efforts focused on isolation of this heat-labile, trypsin-sensitive factor have revealed that it co-fractionates with the major parasite surface glycoproteins and elutes from an anion exchange column in a sharp peak with ,200 mM NaCl (A. Mott  Human foreskin fibroblasts were treated with medium or 5 ng/ml TGF-ß1 for 3 hours in the presence and absence of T. cruzi PCM. TGF-ß1 treatment resulted in a $2-fold up-regulation of 298 unique genes (represented in 309 probe-sets) as compared to cells treated with medium. Genes highlighted above represent the 59 annotated, non-redundant TGF-ß-inducible genes for which expression is reduced by 1.7-fold or greater in the presence of T. cruzi PCM in 3 independent experiments. * Values represent the ratio of the normalized log 2 intensities for treatment (TGF-ß or TGF-ß/PCM)/mock treated controls. doi:10.1371/journal.pone.0023482.t001  unpublished observations) and is likely a component of previously described small shed vesicles [64]. It will be of great interest to identify the molecular basis for the parasite-dependent repression to better assess its functional properties and to probe its potential as a novel antagonist of fibrogenic gene expression and potential immunosuppressive activity in vitro and in vivo.
Overall, this study highlights the ability of a trypomastigoteassociated soluble factor to dampen basal and agonist-induced expression of a key regulator of ECM synthesis as well as genes involved in wound repair and host response to infection. As studies have clearly demonstrated, T. cruzi infection eventually elicits a robust inflammatory response in the host where elevated ECM synthesis and fibrosis are known to contribute to the pathophysiology of Chagas' disease (reviewed in [65]). The cellular responses to T. cruzi released/secreted factors studied here are immediate/ early in nature and provide insights into potential processes that would be expected to occur at early time points of infection, where the parasite-mediated inhibition of host cell fibrogenic gene expression may serve a critical function in facilitating dissemination from the initial sites of infection to the circulation and peripheral organs during acute stages of infection.

Cell Culture and Parasite Maintenance
Human foreskin fibroblast (HFF) BJ cells were acquired from the American Tissue Culture Collection (Manassas, VA, USA) and maintained in DMEM +10% fetal bovine serum (FBS), 1% penstrep, and 2 mM glutamine at 37uC in 5% CO 2 . Tissue culturederived Trypanosoma cruzi trypomastigotes (Y strain) were generated by weekly passage in confluent monolayers of LLcMK 2 cells in DMEM containing 2% FBS as described [5]. Trypomastigotes harvested from cell culture supernatants were washed 3 times in cold serum-free DMEM (SF-DMEM), resuspended in warm SF-DMEM at a concentration of 5610 7 to 1610 8 parasites per ml and incubated at 37uC in 5% CO 2 for 14-18 hours. The resulting 'parasite-conditioned medium' (PCM) was clarified by centrifugation of parasites (1000xg) followed by passage through a 0.2 mm membrane. Prior to treatment of cells, PCM is adjusted to pH 7.4 by addition of a small volume of 1M Tris pH 8.0 and supplemented with heat-inactivated FBS (95uC, 15 minutes) to achieve a final concentration of 2%.

Quantitative Real-Time PCR
HFF were seeded in 6-well plates and grown to 80-90% confluence for 48 hours prior to treatment of cells as indicated. Following treatment, cells were washed with cold PBS and total cellular RNA was harvested using the RNeasy RNA isolation system (Qiagen, Valencia, CA, USA) according to manufacturer's instructions. 200 ng of DNaseI-treated RNA (Invitrogen, Carlsbad, CA, USA) was transcribed to cDNA using the iScript Reverse Transcription system (BioRad, Hercules, CA, USA). Multiplex quantitative real-time PCR was carried out using an ABI 7300 instrument. The amount of experimental cDNA was normalized to GAPDH levels present in the same sample. For quantification of CTGF/CCN2 mRNA each reaction included 20 nM of each primer (forward 59-CTGCCCTCGCGGCTTA-39 and reverse 59GGACCAGGCAGTTGGC TCTA-39) and 10 nM FAM-labeled CTGF/CCN2 probe (6FAM-ACACGTTTG GCCCA-GACCCAACTATG-TAMRA). Primers and probes specific for

Western blot Analysis
Relative expression of CTGF/CCN2 was determined in HFF that were serum-starved for 18 hours prior to treatment with 5 ng/ml TGF-ß for 24 hours in the presence or absence of PCM. Cells were rinsed in cold PBS, lysed directly in 2X Laemmli Buffer containing 5% ß-mercaptoethanol (BioRad, Hercules, CA, USA) and western blot analyses were performed. For MAP kinase activation, HFF were serum starved for 18 hours prior to treatment with 10 mM LPA or 5 ng/ml TGF-ß1 with or without PCM for 5 or 15 minutes. Anti-ß-actin antibodies were purchased from Sigma (St. Louis, MO, USA); anti-CTGF from Santa Cruz Biotechnology (Santa Cruz, CA, USA); anti-p44/42, antiphospho-p44/p42, anti-p38, anti-phospho-p38, anti-SAPK/JNK and anti-phospho-SAPK/JNK from Cell Signaling Technology (Danvers, MA, USA). The ECL Plus chemiluminescence kit (Amersham Pharmacia, Arlington Heights, IL, USA) was used for detection. Blots were scanned and band intensity was quantified using the software program UN-SCAN-IT gel (Silk Scientific Corporation, Orem, UT, USA).

Transient Transfection and Reporter Assays
HFF were seeded in 6-well plates and grown to 50% confluence for 24 hours. Cells were transfected with 1 mg of experimental plasmid DNA and 0.05 mg of control plasmid by lipofection (FuGENE 6, Roche, Indianapolis, IN, USA) for 3-4 hours in DMEM +10% FBS. Cells were washed and incubated in DMEM +10% FBS for an additional 12-14 hours prior to incubation for 6 hours with DMEM-2, PCM prepared in DMEM-2, 5 ng/ml of TGF-ß1 in the presence and absence of PCM. Promoter/reporter constructs used contained ctgf/ccn2 promoter fragments spanning nucleotides 2805 to +17 (2805), 2244 to +17 (2244), 286 to +17 (286) and a construct identical to 805 with a mutation in the Ets binding site (2805 ETS) corresponding to a GGAAT to TCCCG change in the consensus Ets binding motif located in the transcription enhancing factor (TEF) binding element located between 291 and 284. All constructs were kindly provided by Dr. A. Leask [21,27]. Transfection efficiency was measured by cotransfection with pSV-ß-Galactosidase plasmid. Cells were harvested and SEAP and ß-galactosidase assays were performed using the Luminescent ß-galactosidase Reporter System or Great EscAPe SEAP Chemiluminescence Kit (Clontech, Mountain View, CA, USA) according to manufacturer's instructions.

DNA Microarray Hybridization and Analysis
HFF were mock-treated or stimulated with TGF-ß for 3 hours, in the presence or absence of T. cruzi PCM, rinsed 3 times with PBS. RNA was isolated with Trizol reagent (Invitrogen, Carlsbad, CA, USA) and concentrated using an RNeasy kit (Qiagen, Valencia, CA, USA). RNA integrity was evaluated employing an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA), three independent samples for each treatment were processed and hybridized to individual HG_U133 plus 2.0 Affymetrix chips at the Harvard Medical School Biopolymers Facility. Arrays were scanned using a GeneChipH Scanner 3000 (Affymetrix, Santa Clara, CA, USA) and data files containing the unprocessed intensity values were imported into Rosetta Resolver Biosoftware 7.0. Data were pre-processed to reduce systematic errors (i.e. background subtraction and intra-array normalization). Through the application of the Affymetrix-specific error model within Rosetta Resolver [66] the data were transformed into profiles (i.e. scanned, imaged and normalized expression data from a microarray as described in http://www.rosettabio.com/tech/Data_processing_and_analysis_methods.pdf). Intensity ratios were calculated from the statistical combination of replicate profiles in order to increase the confidence in measurements and to obtain fold change values and associated p-values, using the control condition (cells treated with culture medium) as a baseline. Transcripts showing $2-fold change (p,0.01) with respect to matched controls were used to create gene lists and for analysis of biological networks using Ingenuity Pathway Analysis TM software (http://www.ingenuity.com). In accordance to the minimum information about microarray experiment (MIAME) guidelines, the complete raw and processed data files for each array are publicly available at the Gene Expression Omnibus (GEO) database repository (http://www.ncbi.nlm.nih. gov/geo) under accession number GSE16416.

Supporting Information
Table S1 Modulation of fibroblast gene expression by a secreted/released T. cruzi fraction. Human foreskin fibroblasts (HFF) were mock-treated or incubated with cell-free T. cruzi parasite-conditioned medium (PCM) for 3 hours and the isolated RNA was processed and used to hybridize HG_U133 plus 2.0 Affymetrix arrays. Ratios of the normalized log 2 intensities for PCM vs medium were used to calculate the fold-change in HFF gene expression for each array and values from 3 biological replicates were averaged. HFF genes for which the transcript abundance changes $2-fold in either direction are included. (XLSX) Table S2 Inhibition of TGF-ß-inducible gene expression by T. cruzi PCM. HFF were mock-treated or stimulated with 5 ng/ml TGF-ß1 for 3 hours in the presence or absence of T. cruzi PCM and cells were processed for microarray hybridization as described. Ratios of the normalized log 2 intensities for TGF-ß vs medium and TGF-ß+PCM vs medium were used to calculate the fold-decrease in TGF-ß inducible gene expression that was observed in the presence of T. cruzi PCM. Only the non-redundant, annotated genes are shown with the exception of genes for which duplicates showed marked differences in the fold-change values. (XLSX) Table S3 Canonical pathways enriched in the T. cruzi PCM-sensitive subset of TGF-ß inducible genes. TGF-ßinducible genes sensitive to T. cruzi PCM (59 genes from Table 1) were analyzed using Ingenuity Pathway Analysis software. Canonical pathways were identified from the Ingenuity Pathways Analysis library of canonical pathways that were most significant to the data set and associated with a canonical pathway in Ingenuity's Knowledge Base. The significance of the association between the data set and the canonical pathway was measured in 2 ways: 1) A ratio of the number of molecules from the data set that map to the pathway divided by the total number of molecules that map to the canonical pathway is displayed. 2) Fisher's exact test was used to calculate a p-value determining the probability that the association between the genes in the dataset and the canonical pathway is explained by chance alone. (XLS) Table S4 Canonical pathways enriched in the subset of TGF-ß inducible genes that is insensitive to T. cruzi PCM. TGF-ß-inducible genes that were insensitive to T. cruzi PCM (192 genes; Table S2) were analyzed using Ingenuity Pathway Analysis software and canonical pathways were identified from the Ingenuity Pathways Analysis library as described for Table S3.  (XLS)