Hypoxia Supports Epicardial Cell Differentiation in Vascular Smooth Muscle Cells through the Activation of the TGFβ Pathway

Epicardium-derived cells (EPDCs) are an important pool of multipotent cardiovascular progenitor cells. Through epithelial-to-mesenchymal-transition (EMT), EPDCs invade the subepicardium and myocardium and further differentiate into several cell types required for coronary vessel formation. We previously showed that epicardial hypoxia inducible factor (HIF) signaling mediates the invasion of vascular precursor cells critical for patterning the coronary vasculature. Here, we examine the regulatory role of hypoxia (1% oxygen) on EPDC differentiation into vascular smooth muscle cells (VSMCs). Results: Hypoxia stimulates EMT and enhances expression of several VSMC markers in mouse epicardial cell cultures. This stimulation is specifically blocked by inhibiting transforming growth factor-beta (TGFβ) receptor I. Further analyses indicated that hypoxia increases the expression level of TGFβ-1 ligand and phosphorylation of TGFβ receptor II, suggesting an indispensable role of the TGFβ pathway in hypoxia-stimulated VSMC differentiation. We further demonstrate that the non-canonical RhoA/Rho kinase (ROCK) pathway acts as the main downstream effector of TGFβ to modulate hypoxia’s effect on VSMC differentiation. Conclusion: Our results reveal a novel role of epicardial HIF in mediating coronary vasculogenesis by promoting their differentiation into VSMCs through noncanonical TGFβ signaling. These data elucidate that patterning of the coronary vasculature is influenced by epicardial hypoxic signals.


Introduction
Mesenchyme production from the epicardium is a critical developmental event during the generation of coronary vessels [1]. While the origin of coronary endothelial cells has been debated over the years, the consensus is that they arise from a heterogeneous lineage including the sinus venosus, epicardium, neural crest cells and the endocardium [2][3][4][5]. A heterogeneous lineage of the coronary vascular smooth muscle cells (VSMCs) is also expected, however it has been clearly shown that a subset of epicardium-derived cells (EPDCs) originate in the embryonic proepicardial serosa region, undergo epithelial-to-mesenchymal transition (EMT), invade the epicardial connective tissue and eventually the subcompact zone of the myocardium [6][7][8][9][10]. EPDCs are considered an important J. Cardiovasc. Dev. Dis. 2018, 5, 19 2 of 16 multipotent progenitor population that is able to generate these various cell lineages associated with coronary vessels [11,12]. In adults, the epicardial cells can also differentiate into various cell types including fibroblasts, endothelial, cardiomyocyte and VSMCs [13][14][15][16][17][18][19].
While many signaling pathways are involved in the development of the coronary vasculature, we and others have shown transforming growth factor-beta (TGFβ) is required for the proper development of coronary vessels by controlling EMT and further differentiation of epicardial cells into VSMCs [20][21][22][23][24][25][26][27][28]. TGFβ1-3 are a set of pleiotropic cytokines involved in the proliferation, differentiation, migration and EMT of various cell types and the signaling is transduced to the nucleus through the formation of heterodimeric receptor complexes between Type I [TβRI, also termed activin receptor-like kinase-5 (ALK-5)] and Type II (TβRII) II serine/threonine kinase receptors [29]. ALK5 phosphorylates and thereby activates specific SMADs (homologues of Caenorhabditis elegans protein Sma and the Mad (mothers against decapentaplegic) Drosophila protein), which then form heteromeric nuclear complexes with SMAD4. The resulting complexes regulate target genes ultimately serving as canonical transcriptional regulators of TGFβ. The response of this pathway to environmental signals likely contributes to the spatiotemporal specificity of coronary differentiation.
Tissue hypoxia has been shown in a range of contexts to drive adaptive gene expression. Many cellular adaptations to hypoxia during normal mammal embryogenesis are mediated by Hypoxia Inducible Factors (HIFs) which promote the transcription of multiple genes including those that support angiogenesis, energy metabolism and red blood cell production [30,31]. HIFs are heterodimeric basic-helix-loop-helix-Per-Arnt-Sim (bHLH-PAS) transcription factors containing oxygen-sensitive α-subunits that are readily degraded in the presence of oxygen (O 2 ). Once stabilized under hypoxia, HIF-α subunits translocate to the nucleus and dimerize with its β-subunit which then activate multiple global and tissue-specific gene targets [31]. Our laboratories reported that particular cardiac regions (e.g., atrioventricular junction [AVJ] and interventricular septum [IVS]) where major coronary vessels emerge, express nuclear-localized HIF-1α and are highly hypoxic [32][33][34]. Alteration of HIF-1α expression by changing environmental oxygen levels or inducing expression of constitutively active HIF-1α (caHIF-1α) impaired coronary vessel patterning and created a variety of coronary anomalies some of which resemble those observed in patients [35,36]. Increased or decreased expression of Cited2, a regulator of HIF-1α caused abnormal coronary vasculature patterning and permeability [37,38]. We have also shown that forced expression of caHIF-1α in the epicardium of avian embryos resulted in reduced invasion of these cells into the myocardium due to the increased expression of the antagonistic receptor to Vascular Endothelial Growth Factor (VEGF), VEGFR1, which inhibits VEGF signaling [39]. These findings support multiple roles for HIF-1α in coronary development, homeostasis and diseases requiring further investigation.
Because VSMCs are critical for the growth, remodeling and homeostasis of vessels and also involved in cardiovascular pathogenesis (reviewed in [15]), it is important to reveal the functional mediators of epicardial mobilization and the steps involved in their differentiation in the context of the developing epicardium. It was recently reported that hypoxia induces the developmental differentiation of Tbx18+ epicardial cells to VSMCs through Snail [40]. As molecular hypoxic signals are key regulators in differentiation of various stem cell types, we postulated that hypoxia influences the development of the coronary vasculature in part by controlling the plasticity of epicardial cells. In the present study, we sought to identify alternate mechanisms by which hypoxia promotes their differentiation focusing on TGFβ that is strongly implicated in epicardial EMT and differentiation. We previously published that TGFβ-1 or TGFβ-2 induces EMT and smooth muscle differentiation in epicardial cells [21]. In addition, activation of TGFβRIII accesses the Par6/Smurf1/RhoA pathway to mediate epicardial cell invasion [28,41]. Inhibition of p160 rho-kinase (p160RhoK) and rhoA blocks EMT and prevents the appearance of calponin and SMA-positive cells. Our initial studies showed that the specific differentiation of epicardial cells with TGFβ1 into VSMCs was not disrupted by the forced expression of caHIF-1α [39]. Herein, we demonstrate that hypoxia influences RhoA/ROCK through TGFβ signaling to stimulate the differentiation of EDPCs into vascular smooth muscle cells. These observations provide a novel link between microenvironmental and growth factor mediated signals to regulate coronary vascular differentiation. Revealing the mechanisms controlling coronary development could direct the design of new diagnostic therapies and treatments related to cardiovascular anomalies and diseases.

Transfection and Virus Infection
Contends Cells were cultured on collagen I coated 35 mm dishes at 37 • C and transfected with 0.1 nmol siRNA (Invitrogen) and 8 Ml Xtreme siRNA transfection reagent (Roche, Pleasanton, CA, USA). Medium was changed after 24 h before proceeding with other treatments.
Prior to adenovirus infection, cells were plated with Opti-MEM I (Gibco, Thermo Fisher Scientific, Gaithersburg, MD, USA) reduced serum medium containing adenovirus (1 × 10 9 PFU/mL GFP control (Ad-GFP) or dominant negative (Ad-dnRhoA) co-expressing GFP) overnight at 37 • C. Cells were then transferred to 5% FBS DMEM medium and further cultured under hypoxic conditions before immunohistochemical analysis. Similarly, cells were incubated with serum-reduced medium containing polybrene and lentiviral particles overnight and changed to 5% FBS DMEM before further analysis.

Western Blotting
Equal amounts of protein samples quantified by a Bradford assay (Pierce, Thermo Fischer Scientific, Gaithersburg, MD, USA) were loaded onto 10% SDS-PAGE gels (Bio-Rad Laboratories, Hercules, CA, USA) and transferred to a PVDF membrane (Millipore Sigma, Burlington, MA, USA). Membranes were blocked for 1 h at room temperature with 5% BSA in followed by incubation with primary antibodies overnight at 4 • C. Primary antibodies against the following proteins were

Real Time RT-PCR
cDNA was synthesized from total RNA isolated from cultured epicardial cells. Genes were amplified with Universal SYBR Green Master (Roche) by Real-time PCR (Applied Byosystems StepOne Plus, Thermo Fisher) and quantified by the ∆∆ C T method using 18S as the internal control, as previously described [39]. Primer sequences are listed in Table 1.

Statistical Analyses
All experiments were executed in triplicate. Data are represented as mean ± S.E.M. Statistical analyses were performed using student's paired t test. p value < 0.05 was considered significant.

Results
Differentiation of many cell types, including epicardial cells, requires epithelial-mesenchymal transformation (EMT). Since we previously reported that the forced expression of activated HIF-1α ex vivo or in ovo in the epicardium of developing avian embryos promoted epicardial EMT, we used a well-established immortalized mouse EPDC culture system to closely examine the contribution of hypoxia during these developmental events. We first examined the effects of hypoxia on EMT, expecting the loss of epithelial characteristics and the acquisition of mesenchymal cell properties including the loss or redistribution of proteins associated with cell-cell junctions. In comparison to control EPDC cultures that maintain a compact, cobblestone epithelial-like appearance, cultures exposed to hypoxic conditions at 1% oxygen (O 2 ) exhibited an elongated cellular morphology at 24 h.  Figure 1E). Consistent with transcriptional changes, SM22α and SMA proteins in EPDCs were detected following 2 h of 1% O 2 treatment ( Figure 1F). In contrast, we observed no hypoxic effect on the expression of periostin, shown to be highly expressed in EPDCs and in cardiac fibroblasts ( Figure 1C,D) [42].
Observing that hypoxia induces EMT and differentiation of EPDCs, we sought to determine the direct effects of HIF signaling on this process. We used our SM22α-lacZ mouse reporter EPDC line that serves as a model for differentiation [21]. To abrogate HIF's transcriptional activity, we knocked down HIF-1α by infection of lentivirus expressing short hairpin RNA for HIF-1α (shRNA, gift from Dr. Yu-Chung Yang) or control-scrambled shRNA ( Figure 1H). At 24 h of 1% O 2 treatment we did not observe any cell death in either cultures. However, silencing HIF-1α reduced the expression of smooth muscle cell markers in EPDC cultures as SM22-lacZ reporter ( Figure 1I-L) or of α-SMA expression was reduced ( Figure 1K,L). These results imply that hypoxia promotes the differentiation of epicardial cells through the transcriptional activity of HIF-1α. We had previously shown that TGFβ-induced EMT and VSMC differentiation in EPDCs required the downstream kinase activity of ALK5 [21]. Treatment of EPDCs with 250 pM TGFβ1, hypoxic treatment, or both resulted in increased numbers of SMA + and SM22-LacZ + cells ( Figure 1A-D and Figure 2H-K). Selective inhibition of transforming growth factor-beta (TGF-β) superfamily type I activin receptor-like kinase (ALK) receptors [SB: 2.5 µM SB431542] blocked the differentiation of EPDCs in all conditions ( Figure 2E-G,L-N). To quantify the differentiation of EPDCs, we measured LacZ expression in Sm22α-lacZ immortalized epicardial cells. Relative to control normoxic EPDCs, LacZ expression is 5-and 3-fold greater in TGFβ1 or hypoxic treated cultures, respectively. Addition of ALK5 kinase inhibitor (SB431542) impaired the ability of TGFβ1 or hypoxic treated epicardial cells to induce Sm22-LacZ expression ( Figure 2O). These findings support that the ability of hypoxia to promote VSMC differentiation is mediated via the TGFβ pathway. Since hypoxia is known to induce the secretion of cytokines and growth factors in various cell types we next sought to determine its effects on key components of the TGFβ pathway. Transcript levels for TGFβ-1 and -2 and TGFβR3 are increased in epicardial cultures following 24 h of hypoxic treatment ( Figure 2P). Conditioned media collected from EPDCs cultured for 8 and 24 h at 1% O 2 had increased levels of total TGFβ-1 as measured by ELISA ( Figure 2Q). Furthermore, hypoxia induced the phosphorylation of TβRII, previously shown to be required and sufficient to drive EMT and VSMC differentiation in epicardial cells ( Figure 2R).
TGFβ signaling is also known to mount multiple responses. In fact, non-SMAD mediated pathways are associated with TGFβ's ability to induce EMT [43][44][45]. By Western blotting, we observed that hypoxia induced and sustained pERK and pAKT activation for 24 h in EPDCs ( Figure 3A). However, treatment with specific inhibitors for PI3K/AKT (LY294002), p38 MAPK (SB202190), or MEK1/2 (UO126) pathways did not block the hypoxia induced expression of Sm22-LacZ in these epicardial cultures ( Figure 3E,F). We next sought to examine the effects of hypoxia downstream of TGFβ receptor activation by examining the activation of SMAD transcription factors. As transcription factors, SMADs have been shown to promote the differentiation of neural crest cells into VSMCs [46]. Canonical signaling by Alk5 involves SMAD2 and SMAD3 and results in their phosphorylation and nuclear translocation. As a control, we observe that TGFβ-1 stimulates the nuclear translocation of SMAD2 ( Figure 3B) and knockdown of either SMAD2 or SMAD4 inhibits the differentiation of epicardial cells ( Figure 3A). Thirty min after 1% O 2 treatment, phospho-SMAD2 levels varied over a 24 h period ( Figure 3B). To examine if SMAD signaling is required for epicardial transformation into smooth muscle cells, we transfected immortalized Sm22-lacZ epithelial cells with interfering small double-stranded RNA to knock down either SMAD2 or SMAD4 ( Figure 3C). We found that knocking down SMAD signaling inhibited the ability of TGFβ to induce VSMC differentiation ( Figure 3D) but did not affect the ability of hypoxia to do the same ( Figure 3E).
TGFβ can also induce activation of Rho-like GTPases that, among their various roles, control the dynamic cytoskeletal organization of cells including actin remodeling during EMT. It has been previously shown that RhoA is required for VSMC differentiation from proepicardial cells downstream of TGFβ signaling [47,48]. While total RhoA levels are not altered by 1% O 2 treatment in EPDC cultures ( Figure 4A), it significantly induces the activation of RhoA measured by ELISA ( Figure 4B). Western blot data indicate that unlike ALK5/TGFβR I inhibition (SB), specific blockade of RhoA signaling in EPDCs with ROCK inhibitor (Y) does not reduce the hypoxic induction of phospho-SMAD2 ( Figure 4C). However, inhibition of ROCK (Y) reduces the enhanced expression of SMA by either TGFβ and/or hypoxic treatments, similar to the effects of blocking ALK5/TGFβR I (SB), in contrast to inhibiting AKT/PI3K (LY) ( Figure 4D,E). To confirm the involvement of ROCK, EPDC cultures were infected with a dominant-negative RhoA (Ad-dnRhoA) which hampered hypoxia induced expression of SM22-LacZ or SMA compared to control untreated or green fluorescent protein (Ad-GFP) infected EPDCs ( Figure 4F,G). There results argue that hypoxia influences EMT and differentiation of EPDCs, partially through TGFβ-mediated noncanonical signaling through RhoA/ROCK. phospho-SMAD2 activation by hypoxia in EPDC cultures was inhibited by selectively blocking ALK5/TGFβR I (SB; SB421542) but not with inhibitor to PI3K/AKT (LY: 30 µM LY294002) and found to increase with inhibition to ROCK (Y: 10 µM Y27632) (n = 3) * p < 0.05) (C). In contrast, inhibition of ROCK (Y) blocked SMA expression stimulated by either TGFβ (250 pM) or hypoxia, similar to blocking ALK5/TGFβR I receptor (SB: SB431542) (D). Expression of SMA in hypoxic EPDC cultures was significantly compromised by inhibiting ROCK (Y) but not PI3K/AKT (LY), even in cultures treated with exogenous TGFβ (E). Compared with Ad-GFP treated control cultures, SM22 and SMA smooth muscle cell markers were abolished in EPDCs infected with adenovirus expressing dominant negative RhoA (Ad-dnRhoA) (F). Quantification shows significantly decreased LacZ + cell numbers in Ad-dnRhoA than Ad-GFP control treated hypoxic EPDCs cultures (n = 3) * p < 0.05 (G). N: Normoxia; H: Hypoxia; Bar: 50 µm.

Discussion/Conclusions
During early steps in the development of the coronary vasculature, endothelial tubes connect and form a capillary plexus or anastomosis that eventually transitions into the extensive branching network of the mature coronary vessels. This extensive remodeling requires the envelopment of the coronary endothelial cells by smooth muscle cells. Here, we demonstrate that hypoxia plays a critical role in regulating key steps in the differentiation of epicardial cells into vascular smooth muscle cells. Our findings showed that low oxygen conditions enhanced EMT characteristics of epicardial cell cultures and promoted their differentiation into VSMCs. Our results also revealed an important and novel functional connection between hypoxia and TGFβ-mediated EMT and differentiation of EPDCs into VSMCs through the activation of the RhoA/ROCK pathway independent of SMADs ( Figure 5). The development of coronary vessels during embryogenesis requires regulated epithelialmesenchymal transition, cell migration and differentiation of a subset of epicardial cells. We show that epicardial EMT stimulated by low oxygen tension requires HIF-1α, a primary transcriptional mediator of responses to hypoxia, supporting our previous finding showing that constitutive HIF-1α (AdcaHIF1α) mediated EPDC differentiation and induced EMT in primary embryonic quail epicardial cells while hypoxic treatment promoted the expression of ZO-1 in an immortalized EPDC line [36,39]. We also demonstrated that caHIF1α infected primary EPDCs were capable of differentiating into SMA + cells in response to TGFβ-1 treatment [36,39]. The epicardium has been identified as a hypoxic microenvironment expressing Hif-1α and harboring progenitor cells during embryogenesis and in adults [39,[49][50][51]. Supporting the present observations, Jing et al. [40] recently showed that hypoxia induced the differentiation of TBX18 + epicardial cells to undergo EMT and differentiate into VSMCs through HIF-mediated activation of Snail (a major regulator of EMT) that is also known to be induced by TGFβ [52][53][54]. However, in vivo loss of epicardial Snail1 did not lead to any aberrant morphogenetic processes [55]. In the avian epicardium specifically, we showed that Snail over-expression induced epicardial EMT in vitro and enhanced EPDC invasion into the myocardium in vivo [56]. We also demonstrated that HIF-1α promotes the expression of another EMT-related transcription factor, Twist2, associated with the downregulation of E-cadherin that is important to allow epicardial EMT [39]. Thus, hypoxia and TGFβ may control the mobilization of various EMT inducing transcription factors.
As it has been shown that TGFβ signals play a central role in inducing mesenchymal transformation of the epicardium, we hypothesized that the influence of hypoxia on EPDC differentiation occurs via this molecular pathway [20,25,57,58]. While hypoxic treatment induced the expression of TGFβ-1, TGFβ-2 and TGFβ-RIII along with the activation of TGFβ-RII in EPDC cultures, the inhibition of ALK5/TGFβR I suppressed hypoxia-mediated SM22 and SMA induction, consistent with this pathway being required for EPDC differentiation and indicating that this is the primary pathway responsible for EPDC differentiation [23,24].
There are multiple mechanisms by which hypoxia could modulate the specific TGFβ downstream signals to induce differentiation of EPDCs. Hypoxia has been shown to stimulate epithelial to mesenchymal transition (EMT) in other cell types including renal tubular cells [59] and hepatocytes [60] in part by mediating the TGFβ pathway. In response to hypoxia, we observed the phosphorylation of SMAD2 as early as 30 min, comparable to the response to TGFβ treatment of neural crest cells that also differentiate into VSMC [61]. However, inhibition of the canonical TGFβ pathway by knocking down of SMAD-2 or -4 had no effect on the ability of hypoxia to induce differentiation of EPDCs. While loss of TGFβ-RIII accompanies coronary vessel anomalies, we recently identified dysregulation of NF-kB signaling in TGFβ-RIII knockout epicardial cells [23,24]. Interestingly, important crosstalk between HIF and NF-kB signaling previously demonstrated in immune cells may relate to the increased TGFR3 transcript levels we observed in our hypoxic epicardial cell cultures [62].
EMT induction by hypoxia is consistent with previous data indicating that TGFβ-induced EMT and VSMC differentiation in EPDCs was dependent on RhoA [21,48]. In EPDC cultures, hypoxia did not change the total expression of RhoA but did induce RhoA activation in EPDCs, similar to what has been observed with TGFβ treatment alone [48,61]. Interestingly, while inhibition of ALK5/TGFβR I blocked the activation of SMAD2 by hypoxia, SMAD2 activation was not affected by the inhibition of ROCK/RhoA under the same conditions. In fact, we observed that inhibition of PI3K/AKT further increased the hypoxic induction of SMAD2. Thus, hypoxia may play singular roles in mediating alternative TGFβ pathways in a cell specific manner. In epicardial cells it appears that it controls the activation of RhoA which is associated with their transformation into coronary VSMCs. It has been demonstrated that hypoxia induces RhoA and its activation and that RhoA and Rock1 genes are direct HIF targets [63][64][65]. Using chemical and genetic approaches to block the downstream RhoA target Rho kinase (ROCK) led to the decreased expression of either hypoxia-induced SMA or SM22 expression in EPDCs, much the same way that inhibition of RhoA blocked the differentiation of neural crests into SMCs [61]. In all, it appears that hypoxic signals tightly control various components of the TGFβ pathway to regulate this developmental process through RhoA/ROCK ( Figure 5). In the present model system, our ex vivo EPDC differentiation system had a significant influence in the promotion of VSCMs. Hypoxia is a specific micro-environmental cue that has a variety of effects on the maintenance and differentiation of cells. However, the multifaceted cardiac milieu influences the controlled differentiation of EPDCs into various other cell types, including cardiac fibroblasts, to support the coronary vasculature crucial for proper cardiac function. Indeed, multiple transcriptional pathways are required for functional epicardial development, which may most effectively work in trans-or cis- [52].
Importantly, HIF is also known to be upregulated in ischemic tissues as a result of reduced oxygen concentrations as in the case of myocardial infarction [66][67][68][69][70]. Emerging data suggests that under conditions reproducing the hypoxic embryonic environment cells in the adult heart may be reactivated to undergo EMT and migrate to incorporate into new coronary vessels to stimulate the revascularization and regeneration of the injured heart [17,[71][72][73][74][75]. Although TGFβ has been shown to induce EMT and cell differentiation in human adult epicardial cells, differentiation was not mediated by RhoA/ROCK [76]. Together, these results support a model whereby hypoxia and TGFβ activate the RhoA/ROCK pathway to promote the differentiation of EPDCs into coronary VSMCs during development but also show the necessity of further studies to examine in more detail how hypoxia orchestrates the molecular signals associated with the behavior of adult epicardial cells that could help in the development of cardiovascular therapies in response to hypoxic signals following cardiac injury.