Translocating transcription factors in fluid shear stress-mediated vascular remodeling and disease

Abstract

Endothelial cells are exposed to fluid shear stress profiles that vary in magnitude, pulsatility, and directionality due to regional variations in blood vessel structure. Laminar flow at physiological levels is atheroprotective; multidirectional or reversing low (disturbed) flow promotes inflammation and disease; and high or low laminar flow promote outward or inward remodeling, respectively. However, our understanding of how endothelial cells discern these different flow profiles and regulate gene expression accordingly is limited. This article reviews recent studies that identify the TGFβ/Smad, Notch, Yap/Taz, and Wnt/β-catenin pathways as important mediators of flow profile- and magnitude-dependent signaling.

Shear stress in the cardiovascular system

The endothelial cells (EC) that line the inner surface of blood vessels are closely regulated by the hemodynamic forces from blood flow including fluid shear stress, the frictional force from blood flow. Shear stress is critically important in development, physiology, and diseases of the vascular system despite its relatively low magnitude compared to contractile forces or wall stretch from blood pressure. Endothelial cells sense and transduce signals from shear stress to regulate various signaling pathways through a variety of mechanical sensors (5, 64). These regulatory mechanisms shape the vasculature during embryonic development and continue to adjust vascular morphology to optimize the delivery of blood to tissues in adult life.

Increased metabolic demand by an individual tissue results in production of metabolites such as adenosine that induce local vasodilation and thus higher blood flow through the tissue; if a tissue is subject to prolonged hypoxia, secretion of factors such as VEGF induces angiogenesis to increase capillary density. These changes, by lowering resistance, result in higher flow through the arteries that feed the tissue, which, if sustained, stimulates artery remodeling to allow more blood flow. Conversely, decreased tissue demand results in vasoconstriction, capillary regression, and decreased diameter of upstream arteries. Early studies in animals used arteriovenous shunts and vessel ligation to alter flow, which demonstrated that lower or higher shear stress induced, respectively, inward or outward remodeling to reduce or expand vessel lumen diameter (23). Later studies identified signaling molecules (mechanotransducers, vasodilators, vasoconstrictors, inflammatory mediators, etc.) that mediate remodeling (15, 27).

A key aspect of all vascular remodeling is the involvement of inflammatory pathways. For example, an increase in flow results in inflammatory activation of the endothelial cells by high fluid shear stress, resulting in recruitment of monocytes (54). These monocytes are a major source of VEGF as well as metalloproteinases and ECM proteins that mediate vessel expansion. Low flow-driven inward remodeling is also an inflammatory process requiring monocyte recruitment (63). These and other results led to the notion of a fluid shear stress setpoint, where shear stress at the optimal level stabilizes the vessel, while higher or lower shear stress activates inflammatory and remodeling pathways (2, 3).

In addition to shear stress magnitude, flow can also vary in pulsatility and direction. Of particular clinical importance, flow patterns in regions of arteries that curve sharply or branch can be low and oscillatory or multidirectional, collectively termed disturbed shear stress (DSS). DSS leads to a permanent state of instability and mild inflammatory activation (26). By itself, this state is benign but in conjunction with other risk factors such as hyperlipidemia, hypertension, and hyperglycemia, leads to atherosclerotic plaque (2, 64). Laminar shear stress (LSS) in the physiological range promotes stable cell-cell junctions, quiescence, elongation, and alignment in the direction of flow and activates anti-inflammatory pathways. By contrast, DSS fails to induce alignment and promotes proliferation, priming of inflammatory pathways, and increased susceptibility to inflammatory stimuli. Thus, LSS is actively atheroprotective whereas DSS promotes atherosclerosis. This review will examine the roles of four inter-related pathways recently implicated in flow profile- and magnitude-dependent signaling, and their roles in vascular remodeling and disease.

TGFβ pathway

This ancient and highly conserved pathway regulates a wide variety of cellular responses in many different settings. Its activation is initiated by a family of homologous ligands that have been grouped into three subfamilies: TGFβs, bone morphogenetic proteins (BMPs), and activins (29). These proteins bind to type I and II transmembrane receptors that contain cytoplasmic serine/threonine kinases, which are activated and phosphorylate receptor-regulated Smads in the C-terminus. Non-kinase type III receptors sometimes facilitate ligand binding and signaling. Smad C-terminal phosphorylation facilitates their binding to Smad4 and nuclear entry of this dimeric complex. This complex directly binds target gene promoters to regulate gene expression. As Smads are continuously shuttled in and out of the nucleus, the level of nuclear accumulation is correlated to their activation (i.e., ability to induce target gene transcription). Disruption of the genes for TGFβ family ligands, their receptors, and downstream Smads in mice leads to severe and usually lethal vascular defects (6). While much is known about the components involved in signaling, this pathway is highly context-dependent, requiring specific combinations of ligands and receptors for activation (17).

During development, TGFβ stabilizes vessels by inhibiting endothelial cell proliferation and promoting smooth muscle cell differentiation and basement membrane assembly (10). This is also the case in the adult vasculature. However, under some circumstances, both pre-and postnatal, TGFβ can promote Endothelial-to-Mesenchymal Transition (EndoMT) through Smad2/3 activation (58). In these cases, the endothelial cells up-regulate mesenchymal genes and become more migratory. This process is important in both developmental and postnatal angiogenesis, and developmental formation of heart valves and the coronary circulation. EndoMT also plays a role in atherosclerosis where inflammatory cytokines (IFN-γ, TNF-α, or IL-1β) increase expression of TGFβ receptors, thereby sensitizing cells to TGFβ, resulting in Smad2/3 activation. Endothelial cells under these conditions down-regulate endothelial markers and up-regulate mesenchymal markers. DSS of low magnitude also functions as an inflammatory stimulus in this context, promoting Smad2/3 nuclear translocation. In vivo, atheroprone endothelium also showed elevated P-Smad2/3 in mice and blocking this pathway strongly inhibited disease (8).

Whereas TGFβ mainly promotes nuclear localization and subsequent activation of Smad2 and 3, the BMP ligands mainly activate Smads 1, 5, and 8. Despite their homology, Smad1/5/8 and Smad2/3 often induce distinct genes that can play dichotomous biological roles (29). Indeed, recent work suggests that both Smad1/5/8 and Smad2/3 are activated by flow but of different types. In human umbilical vein endothelial cells (HUVECs), Smad1 and 5 were maximally activated by LSS in the physiological range, between 10–20 dynes/cm², with minimal activation by low and/or disturbed flow (3). Physiological flow also promotes a quiescent, anti-inflammatory phenotype, characterized by cell alignment and nuclear exclusion of p65. Interestingly, in human dermal lymphatic endothelial cells (HDLECs), which experience lower shear stress magnitudes, the maxima for NF-kB suppression and cell alignment was shifted to lower shear (~5 dynes/cm²) (3). Thus, different endothelial cell types appear to have different fluid shear stress setpoints.

Genetic evidence also supports a role for BMPs, their receptors, and Smads 1 and 5 in vascular stability. Deletion or inhibition of BMP9 and 10 or their receptors, Alk1 and Endoglin, results in vascular malformations characterized by dilated, leaky, and fragile vessels (21, 56). These effects were potentiated by high shear stress in vivo. In vitro, flow stimulated the association of two BMP receptors, Alk1 and Endoglin, to increase binding of BMPs (4). Taken together, these data lead to a model in which the BMP-Alk1-Smad1/5 pathway promotes vascular stabilization under physiological flow whereas the TGFβ-Alk5-Smad2/3 pathway promotes remodeling under low or disturbed flow.

Notch pathway

The Notch pathway is another highly conserved developmental pathway that regulates a variety of cellular processes. In mammals, it is mediated by the transmembrane receptors Notch 1–4 and transmembrane ligands of the Jagged and Delta families. Binding of a ligand on one cell to a Notch receptor on another cell triggers cleavage of the Notch extracellular domain. This step is permissive for a second, intramembrane cleavage that releases the Notch intracellular domain (NICD), which then translocates to the nucleus where it regulates gene transcription. A hallmark of this pathway is “lateral inhibition,” in which Notch signaling suppresses expression of ligands. This mechanism creates a positive feedback loop whereby a cell expressing higher levels of the ligand inhibits ligand expression in adjacent cells, establishing a stable difference in gene expression. This mechanism plays several roles in vascular biology. During angiogenesis, Notch regulates tip/stalk cell specification, in which tip cells express high levels of Notch ligands, thus activating Notch signaling in the neighboring stalk cells to suppress tip cell fate (38). Notch signaling is also critical in arterial specification as its mutation is associated with the human genetic disease of the vasculature, Alagille Syndrome (50).

Multiple studies have now shown that Notch signaling in endothelial cells is stimulated by shear stress (12, 37, 40, 41, 48, 57). Shear stress triggers cleavage of Notch receptors, release of the NICD, and target gene expression, which can be blocked by a Notch inhibitor, DAPT. Activation of Notch was maximal under high (physiological) shear stress, where it promoted arterial specification (12, 41, 57). This was found to occur through induction of connexin37, which inhibited cell cycle progression, an essential, permissive step in specification (12). Activation of Notch by shear stress in the endothelial cells that line the heart (endocardium) appears to regulate cardiac trabeculation during development (30, 31). Regulation of Notch signaling by flow was also seen in lymphatic endothelial cells, though the effects differed. In this system, laminar flow at a magnitude typical of lymphatics suppressed Notch signaling, which permitted sprouting and lymphangiogenesis (9). In arterial endothelial cells, LSS activated Notch1 signaling and downstream gene expression. Interestingly, Notch1 also polarized to the downstream end of the cell and localized to the nucleus only in aligned and elongated cells; further, knockdown of Notch1 inhibited misalignment. In mice (Notch1^ECKO), EC-specific Notch1 deletion increased vascular inflammation and accelerated atherosclerosis (37).

In all of these studies, Notch activation was associated with stronger cell-cell junctions and vascular stability, which was recently linked to a novel, non-canonical Notch signaling mechanism (48). These authors found that shear stress-induced Notch1 cleavage resulted in production of a functional, transmembrane domain-only fragment. This peptide physically associated with the VE-cadherin transmembrane domain and promoted junctional stability through increased association of VE-cadherin with the transmembrane domain tyrosine phosphatase LAR and its binding partner, TRIO. The effect required the Rac1 GEF activity of TRIO, which activated Rac1 to stabilize cell-cell junctions and reduce permeability.

Yap/Taz pathway

Yap and Taz are transcriptional co-regulators that interpret various mechanical cues (stiffness, stretch, size, and shape) from cell-cell and cell-substrate interactions to regulate proliferation and differentiation (46). Yap/Taz are suppressed by the upstream Hippo signaling pathway, in which the MST1/2 kinases phosphorylate and activate LATS1/2 kinases, which phosphorylate Yap/Taz. Phosphorylation of key serine residues results in Yap/Taz retention in the cytoplasm and degradation. By contrast, dephosphorylated Yap and Taz translocate into the nucleus, bind TEAD factors, and induce downstream gene expression.

In mice, EC-specific deletion of Yap/Taz led to severe vascular defects from impaired growth and sprouting of endothelial cells (24, 62), as well as failure to mediate VEGF responses through transcription (62). In zebrafish, blood flow promoted Yap1 nuclear localization, which was required for maintaining the structure of lumenized vessels (44). They also showed that, in vitro, LSS resulted in Yap nuclear translocation independent of Hippo signaling but dependent on the actin cytoskeleton. These results were consistent with the general notion of Yap as a stimulator of growth and survival in blood endothelial cells.

A second set of studies, however, demonstrated enhanced activation of Yap by DSS associated with inflammatory activation of endothelial cells (60, 61). In HUVECs, Yap S127 phosphorylation increased under LSS and decreased under DSS, indicating Yap activation under disturbed flow. Increased expression of Yap/Taz target genes was also observed under DSS (60, 61). Increased nuclear Yap and reduced P-Yap were also observed in atherosclerosis-prone regions of arteries in mice in vivo. Finally, EC-specific Yap knockdown or over-expression/activation resulted in reduced or increased atherosclerotic plaques, respectively, in hyperlipidemic mice (60, 61). These studies also presented evidence that inflammatory genes are expressed downstream of Yap/Taz. Interestingly, recent studies identified a link between Yap/Taz and JCAD, a gene associated with coronary artery disease. JCAD physically associates with and inhibits LATS1/2, the upstream kinase that inactivates Yap/Taz (22). Low JCAD expression correlated with resistance to disease, consistent with activation of Yap/Taz contributing to atherosclerosis. In lymphatic endothelial cells, however, Yap/Taz is down-regulated under DSS to promote endothelial cell quiescence (53).

Wnt/β-catenin pathway

The canonical Wnt/β-catenin pathway plays diverse but critical roles in development, tissue homeostasis, and disease, most prominently, cancer. β-catenin plays a dual role as both a junctional cytoskeletal linker and a transcriptional co-factor. In unstimulated cells, cytoplasmic β-catenin is rapidly degraded through the multi-component “destruction complex”. Stimulation by Wnt ligand binding to its transmembrane receptors, Frizzled (Fz) and lipoprotein receptor-related protein 5/6 (LRP5/6), inhibits the destruction complex and stabilizes β-catenin. β-catenin consequently accumulates in the nucleus and engages with T-cell factor/lymphoid enhancer-binding factor-1 (TCF/Lef-1) transcription factors to drive target gene expression (36).

In the vasculature, the Wnt/β-catenin pathway plays multiple, distinct roles during development including endothelial cell specification, EndoMT during heart valve morphogenesis, establishment of the blood brain barrier, and angiogenesis (16, 49). Wnt/β-catenin signaling was found to be activated by oscillatory but not laminar flow in blood (14, 34) and lymphatic endothelial cells (7). β-catenin also localized to the nucleus preferentially in atheroprone regions relative to atheroprotective regions of arteries (14). In the majority of these settings (except the brain vasculature), Wnt/β-catenin thus promoted an activated, migratory phenotype. These results are consistent with induction of inflammatory genes by β-catenin (33, 39).

Crosstalk and Integration

It may come as no surprise when one considers that signaling pathways evolved to regulate complex biological processes, but the most striking feature of the pathways discussed above is the extent to which they are integrated in determining biological outputs. Indeed, there is evidence for strong connections between every pair. Notch and Alk1-Smad1/5 pathways cooperatively regulate endothelial cell behavior and gene expression through a variety of mechanisms including co-assembly of NICD and Smad1/5 transcriptional complexes on common target genes (20, 28, 42, 43, 52). Smad2/3 and Wnt/β-catenin also share in transcriptional complexes that regulate a number of common target genes (13, 18, 19, 32, 45, 51, 55). The Yap/Taz pathway directly participates in the destruction complex to regulate β-catenin stability (1). In human embryonic stem cells, Taz directly interacts with Smad2/3-Smad4 complexes and is required for their nuclear accumulation and transcriptional activity. Taz-mediated localization of Smads is downstream of Smad phosphorylation and complex formation, suggesting that Taz aids retention of Smads in the nucleus (59). Smad3, Smad4, and subunits of the NF-κB pathway were shown to coordinately regulate TGFβ-induced transcription (35). A Notch target gene, Nrarp, complexes with the β-catenin-associated transcription factor, Lef1, to enhance expression of downstream genes that regulate vascular stability (47). Of particular interest, a recent paper defined a mechanism through which Yap, Smads, and β-catenin jointly regulate gene expression (11).

It is apparent that these pathways play critical but complex and often divergent roles in vasculature development and remodeling. However, while the data are far from complete, their roles in flow-dependent regulation in adult vasculature appear a bit simpler and an outline of a picture is beginning to emerge. Laminar shear at physiological levels that stabilizes blood vessels activates Notch and Alk1-Smad1/5 pathways whereas activity of Smad2/3, Yap/Taz and Wnt/β-catenin remain low. By contrast, low or disturbed flow associated with remodeling and atherosclerosis activates Smad2/3, Yap/Taz and Wnt/β-catenin. The combined output is induction of inflammatory and remodeling genes, including the mesenchymal genes that promote cell motility and tissue reorganization. To the best of our knowledge, the activity and roles of these pathways in high shear remodeling has not been investigated.

Conclusion

It is clear that endothelial cells can detect and respond differently to flow profiles of different magnitude and directionality (15). Physiological flow within the proper range stabilizes blood vessels; high flow induces outward remodeling; low flow induces inward remodeling; multidirectional or oscillatory flow of low magnitude induces inflammation and is a local risk factor for atherosclerosis. The transcription factors discussed here transmit information from the plasma membrane receptors and cytoskeletal elements that sense forces to the nucleus, thereby regulating gene expression programs (Figure 1). It now appears that NF-κB, Yap/Taz, β-catenin and Smad2/3 have high activity under low flow, where they cooperate to promote remodeling, which includes activation of inflammatory mediators and (to variable extents) induction of EndoMT. These pathways are also activated under disturbed shear and contribute to susceptibility to atherosclerosis. By contrast, Notch and Alk1-Smad1/5 induce genes that contribute to vascular stability; thus, their loss or mutation results in dilated, poorly organized, unstable blood vessels (vascular malformations).

Figure 1: Summary of pathways.

Physiological laminar shear stress activates Alk1/Endoglin (with an unidentified type II receptor denoted RII) to induce Smad1/5/8 phosphorylation and entry into the nucleus to induce expression of target genes that promote vascular stabilization. Laminar shear, possibly acting through integrins, also induces Yap/Taz phosphorylation and cytoplasmic retention/degradation to inhibit signaling. Notch1, in concert with DLL ligands, is stimulated by flow to induce release of the intracellular domain (ICD), which translocates to the nucleus to induce Cx37, which signals to induce expression of the p27 cell cycle inhibitor. By contrast, disturbed shear stress, acting through yet-to-be identified type I, II and III receptors, promotes Smad2/3 phosphorylation. DSS also triggers Yap/Taz dephosphorylation and activation, and inhibits β-Catenin degradation, to drive its nuclear localization, pairing with TCFs and target gene expression.

A great many questions remain to be answered. We know almost nothing about how shear stress regulates Notch, Yap/Taz, receptors for TGFβ family proteins or β-catenin. We also know almost nothing about the mechanisms by which these pathways show precise shear stress magnitude dependence. We have some general concepts about transcriptional complexes but little in the way detailed understanding of how these transcription factors collectively regulate downstream genetic programs. Nevertheless, it begins to appear that vascular stability, remodeling, and disease may follow a set of coherent rules. Understanding the regulation and roles of these translocating transcription factors would be a major step toward unraveling those rules.