Pathways of Egr-1-Mediated Gene Transcription in Vascular Biology

Vascular endothelium forms a continuous cellular interface between the circulating blood elements and the surrounding tissues. Endothelium forms a nonthrombogenic surface and selective permeability barrier capable of modulating vascular reactivity and blood flow. The integrity of the endothelium is fundamental for the maintenance of normal homeostasis. Injury to this lining results in dramatic changes in the functional characteristics of the endothelium, rendering it adhesive and prothrombotic. These changes may result from molecules inducibly expressed or simply secreted by injured endothelium. These initial events are correlated with subsequent inflammatatory and proliferative cellular changes associated with the development of vascular pathology.

In this issue of the American Journal of Pathology, Santiago et al have added another component in the complex chain of molecular events linking endothelial cell injury and growth factor induction. 1 These observations revolve around an inducible transcriptional factor called the “early growth response factor-1” (Egr-1). 2 Evidence is accumulating to suggest that Egr-1 is involved in the regulation of multiple genes within diverse organ systems through protein-protein interactions with protein kinases and other transcription factors. Indeed, Egr-1 and its role in transcription regulation has potential relevance to the pathogenesis of a variety of vascular diseases. Egr-1-mediated gene transcription and its possible role in vascular disease will be reviewed here.

Egr-1, 3 also known as nerve growth factor induced-A (NGFI-A), 4 krox-24, ZIF268, and TIS8, is an 80- to 82-kd protein consisting of 533 amino acids, discovered independently by a number of laboratories searching for factors regulating cell growth and proliferation. 2,5 It is the prototype of a family of zinc-finger transcription factors that includes Egr-2, Egr-3, Egr-4, and NGFI-B. Egr-1 is an example of an “immediate-early response protein” because it is rapidly and transiently induced by a large number of growth factors, cytokines, and injurious stimuli. 2 Egr-1 contains a DNA binding domain consisting of three zinc fingers which are located between amino acids 332 to 416 toward the carboxy-terminal region of the protein. Through these zinc fingers Egr-1 binds specifically to the major groove of DNA at commonly encountered G+C-rich DNA sequences containing the consensus binding code GCG(T/G)GGGCG. Once bound to DNA, Egr-1 alters gene transcription through mechanisms dependent on both coactivators and corepressors. Mutational studies have identified a strong transcription activation zone within the amino-terminal region of the protein between amino acids 1 and 281. 2 Transcriptional coactivators, such as CREB-binding protein (CBP) and p300, can interact directly with the activation region of Egr-1 and increase Egr-1 trans-activation. 6 However, these interactions are relatively weak compared with some sequence-specific transcription factors, such as the p65 (Rel A) component of NF-κB, and their role in authentic Egr-1-mediated gene expression remains unclear. 6

Corepressors such as NGFI-A-binding proteins 1 and 2 (NAB1 and NAB2) negatively regulate Egr-1 activity. NAB1 was identified using a yeast two-hybrid system by its ability to bind a 34-aa inhibitory domain of Egr-1, called R1, located 5′ of the zinc finger binding domain. 7,8 Deletion of R1 results in a marked increase in Egr-1 transcriptional activity and overexpression of NAB1 markedly decreases Egr-1 transcriptional activity. The related protein NAB2 was subsequently discovered because of its strong homology to NAB1. 9 NAB2 functions similarly to NAB1; however, there are important differences between these related proteins. For example, NAB1 is constitutively expressed in most cell types, whereas NAB2 is rapidly and transiently induced by many of the same stimuli that induce Egr-1. Furthermore, the pattern of tissue expression for NAB2 seems to be more tissue-selective than NAB1. Because of these differences NAB2 may play a negative feedback role by down-regulating the burst of Egr-1 activity that accompanies mitogenic, differentiative, or noxious stimuli. Although the mechanism of action of these inhibitors is unknown, it is tempting to speculate that they interfere with Egr-1’s ability to bind coactivators such as CBP or p300.

It has been hypothesized that Egr-1 may play a key regulatory role by linking injurious stimuli to the induction of genes directing the expression of effector molecules that ultimately result in vascular pathology. 10 Egr-1 is inducibly expressed in many different cell types; among the vascular cells known to express Egr-1 are endothelial cells, smooth muscle cells, fibroblasts, and leukocytes. Many stimuli associated with the development of vascular diseases, including shear stress, mechanical injury, platelet-derived growth factors (PDGF), hypoxia, reactive oxygen species, angiotensin II, and acidic fibroblast growth factor (FGF-1), are capable of inducing Egr-1 in tissue culture and, in some cases, in authentic blood vessels. 10,11 Santiago et al have provided compelling evidence to suggest that basic fibroblast growth factor (FGF-2) be included in this list of important mediators regulating Egr-1 expression in endothelial cells following injury. 1,11 They demonstrate that Egr-1 induction by injury involves liberation and paracrine activity of FGF-2 1 (Figure 1) ▶ . This growth factor rapidly activates signal transduction pathways involving the mitogen-activated protein kinases (MAPKs) that converge at the Egr-1 promoter. The authors also show that Egr-1 plays a necessary role in the reparative response of endothelial injury. These signaling events may underlie the pathogenesis of intravascular lesions following conventional coronary intervention in humans. 12

Figure 1.

Model of injury-induced PDGF A-chain gene transcription. Endothelial injury releases FGF-2, which binds to its tyrosine kinase receptor and activates ERK and JNK kinases in a paracrine fashion. These kinases phosphorylate ternary complex factors (TCF), which cooperate with serum response factor (SRF) to induce Egr-1 transcription. Egr-1 can displace Sp1 and other transcription factors from the G+C-rich region of target genes and increase transcription above basal levels through multiple protein-protein interactions with CBP/p300. The basal transcription apparatus (B) is indicated. The number of interactions is related to the number and position of Egr-1 consensus binding sites within the promoter. Other transcription factors (?) may be involved. Activated genes may alter cell growth, migration, adhesion, and survival. For example, the released PDGF A-chain binds to receptors on local cells, stimulating growth and migration of adjacent vascular smooth muscle cells.

Inducible Egr-1 gene expression is mediated through different subgroups of MAPKs, including the extracellular signal-regulated kinase (ERK), c-Jun NH₂-terminal kinase (JNK) (also known as stress-activated protein kinases or SAPKs) and p38 pathways. Once activated, these pathways may lead to an interaction between ternary complex factors (TCF) and serum response factor (SRF) that activates Egr-1 gene transcription by binding to serum response elements (SRE) within its promoter 13,14 (Figure 1) ▶ . Santiago et al have shown that at least some of these events are activated following mechanical injury to cultured endothelium in an endogenous FGF-2-dependent manner. 1 It is possible that protein kinases involved in the induction of Egr-1 transcription may also affect Egr-1’s phosphorylation state and alter its ability to bind other proteins, as well as DNA.

FGF-2’s ability to induce Egr-1 is also noteworthy because it appears to be a factor responsible for a significant portion of Egr-1 production by endothelial cells after mechanical injury. If these findings hold true in authentic blood vessels, FGF-2 may represent a narrow focal point for a therapeutic intervention that breaks the cycle of injury-induced vascular proliferation that leads to lesion formation. However, in vivo, it is possible that other important factors and parallel pathways are involved, with perhaps several stimuli acting in an additive, synergistic, or inhibitory fashion to alter Egr-1 levels beyond those achievable by any single factor alone. In this regard, the role of the Egr-1 corepressors NAB1 and NAB2 in this context remains unknown. NAB2 is present in vascular endothelial cells and smooth muscle cells, and increases in response to injury closely following Egr-1 up-regulation [E.S. Silverman and T. Collins, unpublished data]. The time course suggests that NAB2 plays a role in modulating or counterbalancing the effect of increasing Egr-1 levels.

Following vascular cell activation or injury, Egr-1 is expressed primarily in the nucleus of cells and is capable of altering the transcription of several genes implicated in the pathogenesis of vascular disease, including PDGF-A, 15,16 PDGF-B, 11 FGF-2, 17 apolipoprotein A1, 18 macrophage colony-stimulating factor (M-CSF), 19 TNF-α, 20 tissue factor, 21 urokinase-type plasminogen activator (u-PA), 22 interleukin-2 (IL-2), 23 intracellular adhesion molecule-1 (ICAM-1), 24 copper-zinc superoxide dismutase gene (SOD1), 25 p53, 26 thrombospondin, 27 CD44, 28 and 5-lipoxygenase (5-LO). 29 All of these genes contain one or more Egr-1 consensus binding site within their promoter regions. Because many of these gene products also stimulate the expression of Egr-1, autocrine or paracrine loops within blood vessels are possible. These positive feedback loops serve to amplify and sustain gene transcription through Egr-1-mediated mechanisms. For example, PDGF A-chain and FGF can stimulate expression of Egr-1 and the increased Egr-1 can activate PDGF A-chain or FGF transcription, completing the amplification cycle. 16,30,31

The role of Egr-1 in the regulation of authentic genes in vivo remains to be determined. Most of the studies demonstrating Egr-1 inducibility have involved transient transfection analysis of cells using promoter-reporter constructs of genes suspected of being Egr-1 targets. However, promoter-reporter constructs can behave differently from authentic promoters in the context of other regulatory elements and chromatin. To our knowledge, the only genes definitively linked to Egr-1 expression in vivo have been luteinizing hormone-β (LH-β), tissue factor, and apoprotein A-I genes, based on studies using the Egr-1 knockout mouse developed by J. Milbrandt and colleagues. 32 For example, the human LH-β promoter contains two functional Egr-1 binding sites in the proximal promoter region. In vivo support of these observations is the finding that female homozygous Egr-1 null mice are infertile as a result of luteinizing hormone deficiency. 33 Another example with special relevance from a vascular perspective involves the tissue factor gene. The tissue factor promoter contains a serum response region that binds Egr-1 and activates transcription in vitro in response to hypoxia. 21 In contrast to wild-type mice, mice deficient in Egr-1 fail to produce tissue factor, nor do they deposit fibrin in the pulmonary vasculature under hypoxic conditions. These findings strongly suggest a role for Egr-1 in the transcriptional regulation of tissue factor in authentic blood vessels. 34 The final example of Egr-1’s transcriptional effects in vivo involves the apoprotein A-1 gene. In a mouse model of nephrotic syndrome, levels of apoprotein A-1 in Egr-1 null mice were half those of the wild-type mice. 35 Future studies involving these mice are likely to be particularly useful for identifying the roles of Egr-1 in an authentic biological system. However, attempts to demonstrate Egr-1-mediated transcription may be confounded by redundancy within the Egr family of transcription factors such that, in the absence of Egr-1, changes in gene expression may not be overtly manifested because related transcription factors may substitute and fulfill similar transcriptional roles. Therefore, in the absence of observable changes in transcription, one may not conclude that Egr-1 has no role in the transcriptional regulation of a given gene.

Egr-1’s potential involvement in the pathogenesis of vascular disease was first recognized by Khachigian et al through promoter analysis of the PDGF A-chain and B-chain genes in endothelial cells, and represents the first link between a transcription factor and a target gene in the context of vascular injury. 36 PDGFs are among the most potent mitogens and chemotaxins secreted by endothelial cells and vascular smooth muscle cells. Elevated levels of PDGFs are found in atheroscerotic plaques and are an important link between endothelial injury and the resulting fibroproliferative response that leads to atherosclerosis. 37,38 In an effort to understand the regulation of PDGFs as they relate to the pathogenesis of vascular disease, the promoter sequences were cloned and studied in endothelial cells. The human PDGF A-chain and B-chain gene promoters are highly G+C-rich, contain a TATA box, and have a single transcriptional start site. 39,40 The A-chain promoter is hypersensitive to cleavage by S1 nuclease and contains two overlapping Egr-1 binding sites between −71 and −50 bp from the transcription start site. 41 The B-chain promoter has a cryptic Egr-1 site located −30 to −19 bp from the transcription initiation site. Transient transfection analysis using promoter-reporter constructs containing deletions of the A- and B-chain promoters have determined that the G+C-rich regions are essential for promoter activity and inducibility in endothelial cells. 11,15 Phorbol ester 12-myristate 13-acetate (PMA) was the initial model agonist used to stimulate Egr-1 expression; however, these observations have subsequently been extended to several more pathophysiologically relevant stimuli such as mechanical injury, shear stress, and growth factors. 31,42 These G+C-rich promoter regions are capable of binding recombinant Egr-1 and purified Egr-1 from nuclear extracts of stimulated endothelial cells and vascular smooth muscle cells. Moreover, Egr-1 up-regulation in endothelial cells can significantly activate promoter-reporter constructs above basal levels. Inhibition of Egr-1 or modification of the consensus binding site mitigates promoter activation. These findings have been extended to vascular smooth muscle cells, which are known to be another important source of some PDGFs in blood vessels. 16,43 The role of the Egr-1 transcriptional pathway in authentic blood vessels is less clear. Although mechanical denudation of the endothelial lining of rat aortae leads to increased endothelial and smooth muscle Egr-1 expression before increased PDGF A-chain and B-chain expression, the causal association between Egr-1 and PDGF expression in vivo remains to be definitively proven. 11,16 The list of Egr-1 target genes includes not only growth and coagulation factors, but proteins that could influence growth retardation and cell survival. By stimulating the expression of the gene for TGF-β, Egr-1 may suppress growth of damaged endothelial cells. Additionally, induction of TGF-β may hinder leukocyte recruitment, modulate vascular tone, and increase expression of growth factors such as PDGF. 44 Injury-induced Egr-1 may also activate p53 via the Egr-1 site in its promoter. Induction of p53 genes associated with cell cycle arrest may provide modestly injured endothelial cells with the opportunity to respond to the injury. Lethally injured endothelial cells could be eliminated by induction of p53-dependent apoptotic genes, or by down-regulation of expression of bcl-2. 45 Determining the balance between Egr-1-dependent growth factors and those proteins involved in survival may play a key role in the cellular events associated with vascular injury. Again, future studies using Egr-1 knockout mice or transgenic mice overexpressing Egr-1 within vascular cells may help elucidate the significance of these possibilities in vivo, as they have for LH-β, tissue factor, and apoprotein A1.

The effect of Egr-1 on the enhancement or repression of reporter-gene transcription depends not only on the arrangement of DNA binding motifs within the promoter, but the cell type and nuclear milieu being studied. 5 Possible explanations include: 1) Egr-1’s ability to act in concert with multiple sequence-specific transcription factors; 2) variations in Egr-1’s ability to interact with transcriptional coactivators such as CBP/p300; 3) variations in Egr-1’s ability to interact with transcriptional corepressors such as NAB2; and 4) altered states of phosphorylation that affect Egr-1’s ability to interact with proteins or DNA. It is the sum of all of the transcriptional components and their phosphorylation status that ultimately determines the effect of Egr-1 on a specific promoter.

Egr-1’s ability to interact with other DNA binding proteins has particular relevance to vascular biology. These interactions may occur at one binding site in a gene; they may occur at mutually exclusive binding sites in the gene; or they may occur directly via protein-protein interactions in the absence of DNA. Of all these mechanisms, displacement by Egr-1 at overlapping Egr-1/Sp1 consensus binding sites, also termed Egr-1 displacement of Sp1, has been most thoroughly described in vascular systems 10,11,15 (Figure 1) ▶ . In this regard, Egr-1 shares similar consensus binding sites with transcription factors Sp1 (-GGGCGG-) and Wilm’s tumor suppressor, WT1 (-GNGNGGGNG-). In contrast to Egr-1 and WT1, Sp1 is a ubiquitous zinc finger protein expressed in nearly all cell types and is required for the expression of many essential genes. In general, highest levels of Sp1 expression are found in cells undergoing differentiation, and these high levels of Sp1 may be required for the subsequent induction of tissue-specific genes. 46 WT1 has the properties of a tumor suppressor gene and is expressed predominantly in the kidney and genital organs where it plays an important role in development. 47 Many promoters contain overlapping Egr-1, Sp1, and WT1 consensus binding sites. 5 In vitro studies using recombinant proteins suggest that these transcription factors can displace one another from many promoters, and binding site occupancy is dependent on an equilibrium determined by their concentration within the nucleus and affinity for the binding site. Sp1 is a relatively weak activator of PDGF A-chain and B-chain transcription and is responsible for basal levels of gene transcription in quiescent cultured endothelial cells. 48 When Egr-1 is up-regulated by injury or growth factors it may displace Sp1 from the A-chain promoter and increase transcription. 15 Egr-1’s ability to augment transcription above levels mediated by Sp1 may relate to its ability to interact with the transcriptional coactivator CBP/p300. Unlike Egr-1, Sp1 does not interact with CBP/p300 but directly with components of the basal transcription apparatus. 6 Sp1 may also have a structural role in transcription by maintaining chromatin in an accessible conformation. 46 Regardless of its mechanism of action, Sp1 appears to initiate transcription in a less efficient manner than Egr-1 in endothelial cells. It is conceivable, however, that Egr-1 may not always be an activator of transcription by this mechanism. If CBP/p300 is unavailable to interact with DNA bound Egr-1, due to sequestration or occupancy of binding sites by other transcription factors, the result of Egr-1 displacement of Sp1 could be transcription repression. In contrast to Egr-1 and Sp1, WT1 is a strong inhibitor of PDGF A-chain promoter-reporter gene expression and may function by occupying the G+C-rich promoter region and prevent Egr-1 or Sp1 from binding. 49 However, WT has not been demonstrated to regulate the corresponding gene, and it is unlikely to have a general role in vascular disease given its limited tissue distribution. This raises the possibility that other WT-like transcription factors may exist that play a role in the negative regulation of these genes.

The number and relative positions of the Egr-1 consensus binding sites are also essential determinants of Egr-1’s ability to activate gene transcription. This is true among different genes and within the same gene among different individuals. Studies involving the 5-lipoxygenase gene promoter illustrate this point. The wild-type human 5-LO gene promoter contains 5 tandem Egr-1 consensus binding sites that bind Egr-1 in vitro. 50 In transfection studies, the promoter-reporter constructs respond to overexpression of Egr-1 and these 5 tandem Egr-1 sites are essential for inducibility. 29 An interesting development in Egr-1-mediated transcription was the recent discovery of a family of naturally occurring promoter mutations within the human 5-LO gene by In et al. 51 These mutations are characterized by a variable number of tandem Egr-1 binding sites, from 3 to 6 sites for each allele. The frequency of mutant alleles was quite high in a sample of normal subjects, ranging from 2 to 18% depending on the particular mutation; the wild-type allele was found in 76% of subjects. These promoter mutations are capable of altering Egr-1 binding and reporter gene transcription such that the intensity of binding and degree of trans-activation are proportional to the number of consensus binding sites. 29 It has been hypothesized that this is due to Egr-1’s ability to interact with CBP/p300 through multiple protein-protein contact points 6 (Figure 1) ▶ . The greater the number of binding sites, the more stable is the platform for the recruitment of CBP/p300 and the greater is the effect on transcription activation. The effect of these promoter mutations on the expression of 5-LO in vivo is unknown at this time but is the subject of intense investigation. A similar phenomenon of genetic variability may relate to other promoters with multiple and variably spaced Egr-1 consensus binding sites.

There is ample evidence to suggest that Egr-1 functions in concert with other sequence-specific transcription factors binding to different sites that are appropriately spaced within the promoter; however, their role in vascular biology is less clear. For example, Egr-1 and the steroidogenic factor-1 can bind to different regions of the LH-β subunit gene promoter and synergistically increase transcription. 52 Egr-1 may also interact synergistically with the p65 protein to regulate transcription of the NF-κB1 (p50) gene. 53 Finally, Egr-1 may act synergistically with nuclear factor of activated T cells (NFAT) to augment IL-2 transcription. 54 Because some of these transcription factors are known to interact directly with CBP/p300, it is possible that together with Egr-1 they facilitate transcription by further stabilizing CBP/p300 at the promoter. Direct protein-protein interactions between sequence-specific transcription factors in the absence of DNA are also possible but less well documented. A study by Jain et al suggests that Egr-1 and Sp1 may bind directly to one another depending on Egr-1’s phosphorylation status as regulated by the casein kinase II-dependent phosphorylation pathway. 55 Bound together, they may effectively sequester each other and mitigate binding to gene promoters. 56 Phosphorylated Egr-1 binds less avidly to Sp1, resulting in higher levels of free transcription factor and the ability to interact with promoters.

Following vascular injury, a series of cellular changes takes place in the vessel wall that can result in the development of pathology. These events are preceded by the inducible expression of a series of genes in endothelial cells. Although we have only begun to dissect the signaling pathways activated by vascular injury, it appears that Egr-1 may play a key role in these initial changes. Egr-1 can be activated by multiple mechanisms, including injury-induced release and paracrine activation by FGF-2, as outlined in the paper by Santiago and coworkers. 1 Additionally, Egr-1 can activate several key groups of pathophysiologically relevant target genes, including growth and coagulation factors, cell surface adhesion molecules, and proteins that can alter cell survival. However, induction of Egr-1 and subsequent Egr-1-mediated transcription are complex processes and not a simple “on-off” switch, as sometimes perceived.