TGFβ2-Mediated Production of Hyaluronan is Important for the Induction of Epicardial Cell Differentiation and Invasion

Abstract

In the developing heart, the epicardium is a major source of progenitor cells that contribute to the formation of the coronary vessel system. These epicardial progenitors give rise to the different cellular components of the coronary vasculature by undergoing a number of morphological and physiological changes collectively known as epithelial to mesenchymal transformation (EMT). However, the specific signaling mechanisms that regulate epicardial EMT are yet to be delineated. In this study we investigated the role of TGFβ2 and hyaluronan (HA) during epicardial EMT and how signals from these two molecules are integrated during this important process. Here we show that TGFβ2 induces MEKK3 activation, which in turn promotes ERK1/2 and ERK5 phosphorylation. TGFβ2 also increases Has2 expression and subsequent HA production. Nevertheless, inhibition of MEKK3 kinase activity, silencing of ERK5 or pharmacological disruption of ERK1/2 activation significantly abrogates this response. Thus, TGFβ2 promotes Has2 expression and HA production through a MEKK3/ERK1/2/5-dependent cascade. Furthermore, TGFβ2 is able to induce epicardial cell invasion and differentiation but not proliferation. However, inhibition of MEKK3-dependent pathways, degradation of HA by hyaluronidases or blockade of CD44, significantly impairs the biological response to TGFβ2. Taken together, these findings demonstrate that TGFβ2 activation of MEKK3/ERK1/2/5 signaling modulates Has2 expression and HA production leading to the induction of EMT events. This is an important and novel mechanism showing how TGFβ2 and HA signals are integrated to regulate changes in epicardial cell behavior.

1. Introduction

The coronary vascular system is crucial to provide the necessary oxygen, nutrients and tropic signals to support heart tissue development and function. Thus, malformation or malfunction of the coronary vessels can lead to severe ailments such as congenital defects and coronary artery disease, which affect a significant portion of the population (American Heart Association, 2009) [1]. During embryogenesis, a subset of cells from the outermost layer of the heart or epicardium generates the different components of the coronary vasculature through a complex process known as epithelial to mesenchymal transformation (EMT)[2]. This process involves cell detachment from the epicardial layer, proliferation, differentiation and invasion through the subepicardial space that lays between the epicardium and myocardium [3]. Thus, EMT is a crucial mechanism through which progenitor cells become specified and contribute to the formation of the cardiovasculature. Although many questions remain to be answered in regards to the specific mechanisms that regulate EMT, recent studies have identified transforming growth factor β2 (TGFβ2) as an important regulator of this process in epicardial cells [4]. TGFβ2 is a member of a large family of structurally related cytokines including activins, bone morphogenic proteins (BMPs) and TGFβs. In mammals, the immediate TGFβ family consists of three isoforms: TGFβ1, TGFβ2 and TGFβ3. However, only TGFβ2 is highly expressed in early stages of epicardial and coronary vascular formation, suggesting a predominant role for TGFβ2 in the regulation of epicardial biology in vivo [5].

TGFβ2 transduces its signals by interacting with heteromeric complexes composed of a type I and a type II TGFβ receptor (TBRI and TBRII, respectively), and in some cases, a type III TGFβ (TBRIII) receptor known as betaglycan [6]. However, cellular responses to TGFβ2 are primarily mediated by TBRI and TBRII. Upon binding to TGFβ2, the type II receptor recruits and activates the type I receptor which in turn promotes the activation of downstream signaling molecules, most notably, the Smad transcription factors.

Nevertheless, TGFβ2 signals can also activate members of the MAPK pathway such as p38, JNK and ERK1/2 to induce cellular responses and appear to regulate the expression of hyaluronan synthase 2 (Has2) in highly invasive cells [7–9].

Has2 is one of three hyaluronan synthase isoforms found in mammals (Has1, 2 and 3) and is present in the membrane of most mammalian cell types, where it produces the extracellular matrix component hyaluronan [10]. Gene knockout studies have shown that Has2 is the most relevant Has isoform during cardiovascular development. While Has1 and Has3 knockout animals are viable and fertile, mice lacking Has2 die at embryonic day (E)9.5 due to severe cardiovascular abnormalities [11]. Has2 knockout mice contain virtually no hyaluronan (HA), which prevents the proper formation of the vasculature and endocardial cushions. These alterations in turn cause abnormal blood flow, malformation of the heart valves and septae, and result in premature death.

Because of the early lethality of the Has2 knockout model, it is not known whether Has2 and its product HA are important for developmental events that occur beyond E9.5, such as epicardial and coronary vascular formation. However, evidence from normal embryos and in vitro studies suggests that Has2 and HA may also play a role in later stages of cardiovascular development. For example, just as in the endocardial cushions, HA is highly abundant in the epicardium, around the time when cells start to undergo EMT [12]. Furthermore, HA induces epicardial cell differentiation and invasion in vitro [13].

Despite the relevance of TGFβ2 and Has2 during embryogenesis, it remains unclear whether signals from these molecules interact and become integrated in a developmental context. Thus, the aim of the present study was to determine whether TGFβ2 contributes to the regulation of Has2 and define the potential role of these interactions in the induction of epicardial EMT.

2. Materials and Methods

2.1 Cell culture

The inducible immortalized epicardial cell line was generated and cultured as previously described [4]. Briefly, the hearts of transgenic mouse embryos harboring a thermolabile SV40 TAg gene were harvested at E11.5, placed on collagen culture dishes and incubated for 15 hrs at 37°C to allow for the formation of epicardial monolayers. The hearts were then removed and the resulting monolayers were propagated at 33°C in Dulbecco’s modified Eaglés medium (DMEM) containing 10% fetal bovine serum (FBS), antibiotics, insulin-transferrin-selenium (Invitrogen, Carlsbad, CA) and mouse gamma interferon (10 units/mL, R&D systems, Minneapolis, MN).

2.2 Reagents

Human recombinant TGFβ2 was obtained from R&D Systems and used at a concentration of 2 ng/mL. The hemagglutinin-tagged kinase-inactive MEKK3 construct (K391M) was kindly provided by Dr. Richard Vaillancourt (University of Arizona, Tucson, AZ). The ERK activation inhibitor peptide (500 nM) was purchased from Calbiochem (San Diego, CA) and the CD44 blocking antibody (clone KM201) from Southern Biotech (Birmingham, AL). Antibodies against MEKK3 (H70), pERK5 and Has2 were obtained from Santa Cruz Biotechnology (Santa Cruz, CA) while those against hemagglutinin and β-actin were from Roche Applied Science (Indianapolis, IN) and Affinity Bio Reagents (Rockford, IL), respectively. Antibodies recognizing ERK1/2, ERK5, phospho-p38, p38, phospho SPK/JNK, SPK/JNK and vimentin (R28) were purchased from Cell Signaling Technology (Danvers, MA). Antibodies against phospho-ERK1/2 and phospho-MEKK3 were obtained from Sigma (St Louis, MO) and the Vaillancourt lab, respectively.

2.3 Measurement of HA

Cells were counted, seeded onto 6-well plates and subjected to various treatments. Cell culture supernatants were collected 24 and 48 hrs post treatment and HA concentrations were determined using the Duoset ELISA development system (R&D systems) according to the manufacturer’s instructions. Briefly, 96-well plates were coated with 0.5 μg/mL of recombinant human Aggrecan and incubated overnight at 25°C. Next, plates were washed, blocked with a 5% Tween-PBS solution, and incubated with HA standards or sample supernatants for 2 hrs. Plates were then washed, incubated with 0.3 μg/mL of biotinylated recombinant human Aggrecan for 2 hrs and washed again. Next, horseradish peroxidase-streptavidin was added for 20 min followed by washes. A substrate solution (H₂O₂: tetramethylbenzidine) was then added until sufficient color developed (~15 min) and the reaction was stopped with 2N H₂SO₄. The optical density of each well was determined using a microplate reader set to 450 nm with a wavelength correction of 540 nm to account for optical imperfections in the plate. A linear standard curve was used to calculate HA concentrations and the data are shown as the percentage change in HA production as compared to the untreated controls.

2.4 Western Blotting

Epicardial cells were serum starved overnight and left untreated or treated with various inhibitors and/or TGFβ2. Total cellular lysates were then resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred onto a polyvinylidene fluoride membrane. After blocking in 3% BSA, membranes were probed with different primary and secondary antibodies. Detection was performed using Super Signal West Pico substrate (Pierce). Quantification by densitometry was performed with the ImageJ program (NIH, Bethesda, MD).

2.5 Immunofluorescence

Cells grown and treated on glass coverslips were fixed with paraformaldehyde, permeabilized with Triton X-100, and blocked with BSA. Cells were then incubated with anti-vimentin antibody overnight, washed and fluorescently labeled with AlexaFluor- 594 secondary antibody (Molecular Probes, Eugene, OR) for 1 hour at room temperature. Fluorescent images were taken with the Leica DMLB fluorescence microscope (Leica, Bannockburn, Ill) using the Image ProPlus software (Media Cybernetics, Bethesda, Md).

2.6 Real-time PCR

After the indicated treatments, Total RNA was isolated from cells using the RNA-STAT60 reagent (Tel-test). cDNA synthesis was conducted using Transcriptor First Strand cDNA Synthesis Kit (Roche Applied Science) according to the manufacturer’s instructions. Quantitative real-time PCR was performed as previously described [14]. The gene RPS7 was used for normalization. The following primer sequences were employed: RPS7: AGCACGTGGTCTTCATTGCT, CTGTCAGGGTACGGCTTCTG; Has2: GTTGGAGGTGTTGGAGGAGA, AAAGCCATCCAGTATCTCACG; ERK5: TTGGTCACCACATCAAAAGC, TACGGGGTGGTGTCTTCG. All oligonucleotide primers were synthesized by Integrated DNA Technologies (Coralville, IA). The gene-specific probes were obtained from the Universal Probe Library (Roche Applied Science).

2.7 Proliferation Assays

Cells were seeded into 96-well culture plates and routinely cultured overnight. Next, cells were serum starved for 1 hour, followed by treatment with TGFβ2. Cell proliferation was assessed 24 and 48 hrs post treatment using the Vybrant® MTT cell proliferation assay kit (Molecular Probes) as indicated by the manufacturer.

2.8 Invasion Assays

The ability of TGFβ2 to induce cellular invasion was determined by using a transwell chamber system, as previously described [14]. Briefly, cells fluorescently labeled with Calcein AM were placed onto collagen gels casted in the upper chambers of a transwell 96-well plate. Cells were then treated with different inhibitors and/or TGFβ2 and incubated in low serum media for 24 or 48 hours. Invasion was quantified by measuring fluorescently labeled cells that moved through the collagen gels and crossed into the lower chambers of the transwell system. Fluorescence was determined at 538 nm using the Spectramax Gemini plate reader (Molecular Devices, Sunnyvale, CA).

2.9 siRNA Experiments

Small interfering RNA (siRNA) against ERK5 and control siRNA (siRNA-A) were purchased from Santa Cruz Biotechnology. Cells grown to 50% confluence in 6-well plates were transfected with either ERK5 siRNA or control siRNA using XtremeGene siRNA transfection reagent according to the manufacturer’s instructions (Roche Molecular Systems). Following transfection, cells were incubated for 48 h in a medium containing 10% FBS prior to their use in different experiments.

2.10 Statistical Analysis

All quantitative data were analyzed using two sample Student’s t tests. The data are presented as the means ± S.D. of 2 independent experiments performed in triplicate, except for invasion assays, in which each condition was analyzed in sets of 6. Differences of P<0.05 were considered to be significant.

3. Results

3.1 TGFβ2 induces HA production via MEKK3-dependent pathways

Although TGFβ2 has been shown to induce HA biosynthesis in cancer cells [15], the specific intracellular signals that govern this response or whether this process also occurs in a developmental context remains to be elucidated. Here we investigated whether TGFβ2 is able to regulate HA production in epicardial cells and explored involvement of the MAP3K MEKK3 and its downstream signals, as this molecule is crucial for the proper development of the vasculature [16]. For this purpose, we quantified the presence of HA in the cell culture media following stimulation with TGFβ2 and/or various inhibitors using an ELISA-like method. As shown in figure 1, TGFβ2 induces a significant increase in HA production as early as 24 hrs post-treatment. In contrast, we detect a substantial decrease in HA synthesis when cells are transfected with a kinase-inactive MEKK3 construct (KI-MEKK3) prior to the addition of TGFβ2. Similarly, pretreatment with an ERK1/2 inhibitor peptide or knockdown of ERK5 with siRNA significantly impairs the ability of TGFβ2 to increase HA levels. Ctrl siRNA does not significantly affect HA production following TGFβ2 stimulation. These findings suggest that MEKK3-dependent pathways play an important role in the regulation of HA production by TGFβ2.

Fig. 1.

Regulation of Hyaluronan production by TGFβ2-mediated signals. (A) Cells were treated with TGFβ2 (2 ng/mL) in the presence or absence of KI-MEKK3, ERK1/2 inhibitor peptide, ERK5 siRNA alone, ERK5 siRNA in combination with ERK1/2 inhibitor peptide or control siRNA. Total amounts of HA secreted into culture supernantants were quantitated after 24 and 48 hrs. The results are shown as fold change in HA production as compared to the 24h control. (B) Cells were left untransfected or transfected with KI-MEKK3 and incubated for 48 hrs to verify the expression of the KI-MEKK3 construct. Lysates were subjected to Western Blotting (WB) with anti-hemagglutinin and anti-actin antibodies. (C) Real-time PCR and WB were performed to determine the effect of ERK5 siRNA in silencing ERK5 at the mRNA and protein levels. *, p<0.05 as compared to the untreated control; #, p<0.05 as compared to treatment with TGFβ2 alone.

3.2 TGFβ2 promotes Has2 expression through a mechanism requiring MEKK3

Because HA is primarily synthesized by Has2 in embryonic cells [17], we sought to determine whether TGFβ2 regulates Has2 expression in our epicardial cell model. Here, we performed RT-PCR and Western Blotting to determine the effect of TGFβ2 on Has2 mRNA and protein levels. As shown in figure 2, TGFβ2 significantly increases Has2 mRNA (Fig. 2A) and protein (Fig. 2B) in a time-dependent manner. We detect up to a 6 fold increase in Has2 message at 16 hrs following stimulation with TGFβ2 as well as significant increases in Has2 protein starting at 16 hrs. Next, several components of the MEKK3 pathway were inhibited in order to assess their role in TGFβ2-mediated responses. Transfection with KI-MEKK3, treatment with an ERK1/2 inhibitor peptide, or knockdown of ERK5 with siRNA prior to stimulation with TGFβ2 diminishes Has2 expression. Furthermore, concomitant silencing of ERK5 with pharmacological inhibition of ERK1/2 completely abrogates TGFβ2-mediated increases in Has2 expression. Thus, MEKK1, ERK1/2 and ERK5 are necessary for the regulation of Has2 expression by TGFβ2.

Fig. 2.

Involvement of MEKK3-dependent pathways in the regulation of Has2 expression by TGFβ2. (A) Has2 mRNA levels were detected in cells treated with TGFβ2 for 4–48 hrs in the presence or absence of KI-MEKK3, ERK1/2 inhibitor peptide, ERK5 siRNA alone, or ERK5 siRNA together with ERK1/2 inhibitor peptide. (B) Cells were treated with TGFβ2 (2 ng/mL) for 4–48 hrs and lysates were subjected to WB with anti-Has2 and anti-actin antibodies. The graphs represent WB quantification by densitometry. The results are expressed as Has2 band intensity relative to the actin bands. (C) Cells were transfected with KI-MEKK3 or ERK5 siRNA and/or treated with ERK1/2 inhibitor peptide followed by stimulation with TGFβ2 for 24hrs. WB and densitometry analysis were performed as in 2A. *, p<0.05 as compared to the untreated control; #, p<0.05 as compared to treatment with TGFβ2 alone.

3.3 MEKK3 is activated by TGFβ2 and in turn induces ERK1/2 and ERK5 phosphorylation

A crucial step for MEKK3 activation is the specific phophorylation of the serine residue at position 526 of this protein [18]. Thus we explored whether MEKK3 is phosphorylated at S526 in response to TGFβ2 stimulation. As shown in figure 3A, treatment with 2ng/mL of TGFβ2 significantly induces MEKK3 phosphorylation at S526 in a time-dependent manner.

Fig. 3.

MAPK activation in response to TGFβ2. Cells treated with TGFβ2 for 0–60 min were employed for WB with anti-phospho-MEKK3 (top panel) and reblotted with anti-MEKK3 (bottom panel). Densitometry measurements are expressed as phospho-MEKK3 band intensity relative to the MEKK3 bands. (B) Cells were treated with TGFβ2 for 0–60 min and WBs were performed with antibodies against the phosphorylated and total forms of p38, ERK1/2, SPK/JNK and ERK5. (C) Cells were transfected with KI-MEKK3 and then stimulated with TGFβ2 for 0–60 min. WBs were performed with antibodies against the phosphorylated and total forms of ERK1/2 and ERK5. *, p<0.05 as compared to the untreated control; #, p<0.05 as compared to treatment with TGFβ2 alone.

As part of the MAPK relay cascade, activation of MEKK3 has been shown to induce multiple downstream kinases including p38, ERK1/2, SPK/JNK and ERK5 [19–21]. Thus, we investigated whether any of these molecules are also phosphorylated in response to TGFβ2. We detect that ERK1/2 and ERK5, but not p38 or SPK/JNK are phosphorylated following TGFβ2 stimulation (Fig.3B). This is consistent with our previous results which indicate an involvement for MEKK3, ERK1/2 and ERK5 in TGFβ2-mediated responses. To determine whether MEKK3 is upstream of ERK1/2 and ERK5 following TGFβ2 stimulation, we transfected our epicardial cells with KI-MEKK3 prior to the addition of TGFβ2. As shown in figure 3C, KI-MEKK3 completely abrogates the ability of TGFβ2 to induce ERK1/2 and ERK5 phosphorylation, suggesting that the activation of these molecules in response to TGFβ2 is directly under the control of MEKK3.

3.4 HA and MEKK3 play a role in TGFβ2-mediated cell invasion and differentiation

Depending on the cellular and physiological context, TGFβ2 as well as HA-dependent signaling cascades have been shown to induce cell proliferation, differentiation and invasion [4, 13, 22, 23]. However, it is not known whether these pathways converge or work independently to induce these cellular responses. Furthermore, as we have shown that TGFβ2 induces the MEKK3 pathway and promotes HA production by Has2, we sought to determine whether these molecules are able to regulate epicardial cell biology. As shown in Figure 4A (lane 2), stimulation of epicardial cells with TGFβ2 leads to an increase in the expression of the mesenchymal cell marker vimentin (red) as early as 24hrs post-treatment. This response is decreased in the presence of KI-MEKK3, ERK1/2 inhibitor peptide or ERK5 siRNA alone. (Fig 4A, lanes 3, 4 and 5, respectively). Remarkably, knockdown of ERK5 by siRNA together with pharmacological inhibition of ERK1/2 (lane 6), disrupts the response to TGFβ2 considerably more than when only one of these molecules was inhibited, indicating a synergistic effect between ERK1/2 and ERK5 in the transduction of TGFβ2 signals. Additionally, degradation of HA by hyaluronidase (lane 7) and blockade of the HA receptor CD44 (lane 8) also abrogates the ability of TGFβ2 to increase vimentin expression. Thus, HA production and the integrity of the polymer are important for TGFβ2 to induce morphological changes in epicardial cells.

Fig. 4.

Involvement of HA and MEKK3 in TGFβ2-mediated cellular invasion and differentiation. (A) Cells were left untreated, treated with TGFβ2 alone or TGFβ2 in the presence of KI-MEKK3, ERK1/2 inhibitor peptide, ERK5 siRNA, CD44 blocking antibody or hyaluronidase. Cells were then fixed and immunofluorescently stained with an antibody against vimentin. (B) Fluorescently labeled cells were stimulated as in 4A and incubated for 24 or 48 hrs to allow for invasion into collagen gels and through an 8μm pore membrane (C) Cell proliferation was evaluated following stimulation with TGFβ2 for 24 and 48 hrs using the MTT assay. 10% FBS was used as a positive control. *, p<0.05 as compared to the untreated control; #, p<0.05 as compared to treatment with TGFβ2 alone.

Next, we tested the role of TGFβ2 in epicardial cell motility by using a modified Boyden chamber assay. Here we demonstrate that TGFβ2 induces cellular invasion (Fig. 4B, white bars) and that pretreatment with KI-MEKK3, ERK1/2 inhibitor peptide and/or ERK5 siRNA (gray and stripped bars) significantly diminishes this response. Similarly, pretreatment with hyaluronidase or with CD44 blocking antibody (color bars) effectively abrogates the ability of TGFβ2 to promote cell motility. These observations support a role for MEKK3 and HA-dependent pathways in the regulation of epicardial cell invasion by TGFβ2. This enhanced invasion is not the result of increased cell number, as cells stimulated with TGFβ2 do not significantly proliferate after 24 or 48 hrs (Fig. 4C). Together, our data indicate an important functional connection between TGFβ2 and HA-mediated signals and underscores the role of these interactions in the regulation of epicardial cell biology.

4. Discussion

TGFβ2 is highly expressed in the developing heart and a number of studies have demonstrated a prominent role for this molecule in the regulation of EMT [4, 5]. However, many other proteins, such as Has2 [11] and members of the MAPK pathway [24–26], have also been identified as relevant for the EMT process and how these players may be interconnected remains largely unknown.

Our present study provides the first direct evidence that epicardial cells increase HA production in response to TGFβ2 and that this synthesis is important for the induction of cellular differentiation and invasion. We found that TGFβ2 is able to modulate HA synthesis via upregulation of Has2 expression through a mechanism requiring MEKK3, ERK1/2 and ERK5 (Fig. 5). These results demonstrate a novel functional relationship between TGFβ2, MAPK and HA-mediated signals that is crucial for the orchestration of changes in epicardial cell biology. Also, our data further confirm initial evidence involving each of these molecules in the regulation of cardiovascular morphogenesis.

Fig. 5.

A schematic diagram of integration between the TGFβ2 and HA pathways in epicardial cells. We propose that TGFβ2 induces Has2 expression and HA production through MEKK3-dependent mechanisms. These increased HA levels subsequently induce intracellular signals via the CD44 receptor or alternative mechanisms, leading to cellular differentiation and invasion.

Despite its simple carbohydrate structure, HA has been shown to have profound effects on multiple developing organs and tissues, including the brain, heart and eyes [27–29]. HA not only provides structural support as a component of the extracellular matrix, but it also binds cell surface receptors and induces cellular responses such as differentiation and migration [13, 30, 31]. Although it has been established that these responses are dependent on HA size, much controversy remains in the field in regards to whether it is the high or the lower molecular weight HA that is biologically relevant. HA is synthesized by Has2 as a high molecular weight polymer of up to 2×10⁶ Daltons but, once excreted into the extracellular space, HA can be degraded into smaller oligosaccharides by enzymes known as hyaluronidases [10, 32]. In the context of cardiovascular development, HA oligosaccharides have been shown to induce angiogenesis but inhibit EMT [33, 34]. On the other hand, high molecular weight HA promotes EMT in endocardial cushions and epicardial cells [11, 13]. Thus, it is likely that HA must remain in its native high molecular weight state in order for EMT to take place and its degradation into small fragments may act as a switch that turns off the EMT process. In this regard, our current findings are consistent with previous studies as we show that degradation of native HA by hyaluronidases impairs cell differentiation and significantly decreases the invasive response to TGFβ2 in epicardial cells (Fig. 4). Nevertheless, HA degradation does not completely abrogate the ability of TGFβ2 to induce biological responses, suggesting that other mechanisms that are independent of HA may also mediate TGFβ2 signals. Indeed, multiple proteins have been deemed important for the regulation of EMT by TGFβ2 including p160 rho kinase, Smads, Snail, and the beta-catenin/TCF/Lef pathway [4, 35–37]. Thus, it is likely that HA is only one of several key players that participate in the modulation of TGFβ2-induced cellular responses.

MEKK3 is a mitogen activated protein kinase kinase kinase (MAP3K) and as such it participates in the transduction of extracellular signals by triggering the activation of several downstream kinases. The activation of these MAPK cascades results in a variety of responses, including proliferation, differentiation, migration and apoptosis, depending on the cellular context [38]. The importance of MEKK3 in cardiovascular development has been highlighted by targeted disruption of the MEKK3 gene in mice, which leads to severe vascular and myocardial abnormalities and results in embryonic death by E11 [16]. Furthermore, lack of MEKK3 kinase activity has been shown to inhibit the ability of endocardial cushions to undergo EMT while the presence of a constitutively active form of MEKK3 promotes EMT in areas of the heart that do not normally undergo this process, such as the ventricular endocardium [25]. Thus, our results further validate the role of MEKK3 in developmental EMT and also reveal a previously unknown mechanism of MEKK3 activation through TGFβ2. Interestingly, constitutively active MEKK3 has also been shown to induce TGFβ2 expression [25], suggesting that a positive feedback loop may exist between these two molecules.

Among the proteins that can be activated in response to MEKK3, the MAPKs ERK1/2, p38, JNK and ERK5 have all been shown to play a role in the EMT process [26, 39–41]. However, no studies have explored whether any of these molecules are specifically under the control of MEKK3 during developmental EMT. Thus, we assessed this important question and our findings demonstrate that, in epicardial cells, MEKK3 activates ERK1/2 and ERK5, but not SPK/JNK or p38, in response to TGFβ2 (Fig. 3). Furthermore, ERK1/2 and ERK5 activity are critical for TGFβ2 to induce Has2 expression. In contrast, HA production and cellular responses following TGFβ2 stimulation are only partially inhibited in the absence of ERK1/2 and ERK5 activity. A possible explanation for these results is that TGFβ2 may regulate Has2 activity through several mechanisms and therefore, a blockade of Has2 expression does not necessarily prevent Has2 from being activated in the cell membrane and inducing HA synthesis. In fact, several studies have suggested that posttranslational modifications such as glycosylation and phosphorylation are important for Has activity, although the specific sites for these modifications are yet to be identified [42, 43]. Thus, it is likely that ERK1/2 and ERK5 are required for the transcriptional, but not posttranslational, regulation of Has2 by TGFβ2.

Our current data suggesting a contribution for ERK5 in the development of the cardiovascular system is consistent with in vivo evidence presented in the literature. Disruption of the ERK5 gene in mice, for example, leads to severe and lethal defects such as abnormal heart looping and a disorganized vasculature [44]. Also, although ERK1 and ERK2 knockout mice exhibit only minor non-lethal developmental defects, these proteins have been implicated in the induction of HA and the regulation of multiple cellular responses during embryogenesis [45–48]. Thus, ERK1/2 may have a collaborative and/or synergistic role in the transduction of developmental cues. It is worth noting that previous studies using normal adult and cancer cells have implicated JNK and p38 in TGFβ2-mediated EMT but similar experiments utilizing embryonic cells have failed to identify such connection [4, 7, 49]. These observations, together with our results suggest that TGFβ2 selectively regulates MAPKs based on the specific cell type and that ERK1/2 and ERK5 are relevant in the context of epicardial EMT regulation by TGFβ2.

The present study is the first to our knowledge to reveal a novel MEKK3-dependent mechanism through which TGFβ2 regulates Has2 and its product hyaluronan. This regulation is partially required for the induction of epicardial cell differentiation and invasion and underscores a direct functional connection between TGFβ2 and HA-mediated pathways.