A unifying concept in vascular health and disease

See allHide authors and affiliations

Science  20 Apr 2018:
Vol. 360, Issue 6386, pp. 270-271
DOI: 10.1126/science.aat3470

Not unlike Tolstoy's remark about happy versus unhappy families, current wisdom in vascular biology holds that healthy blood vessels are mostly similar, whereas vessels in different vascular diseases are mostly different. But is this really the case? An evaluation of the literature suggests that unresolved vascular remodeling may be a key element of virtually all vascular diseases. This commonality raises the possibility of unifying principles that govern vascular remodeling and the possibility that methods to restore normal remodeling could effectively treat multiple disease states.

Postnatal physiological vascular remodeling encompasses the sprouting of existing blood vessels to form new ones (angiogenesis), and the growth and remodeling of small vessels into larger arteries that support higher blood flow (arteriogenesis). Importantly, both processes are self-limited. Angiogenesis is stimulated by tissue demand for oxygen and nutrients due to tissue growth, exercise, or wound healing (1). Secretion of vascular endothelial growth factor (VEGF) by cells in ischemic tissue is the main driver, but once blood flow is established in the new vessels, oxygen and nutrient transport into the tissue decreases VEGF expression, shuts down angiogenesis, and restores vascular stability. Arteriogenesis is triggered mainly by increased blood flow and hence fluid shear stress through small vessels. This induces a local inflammatory response, leading to VEGF production (2). The resultant outward remodeling, vessel enlargement, and de novo formation of arterial vasculature restore normal levels of shear stress, which resolves vascular wall inflammation and restores normal function. These two processes are of relatively short duration, on the order of days to weeks in mouse models, and fully restore normalcy.

By contrast, prolonged or chronic stimulation of angiogenesis or arteriogenesis results in pathological remodeling and leads to morphologically and functionally abnormal vessels. One example is atherosclerotic plaques that form as a result of converging biomechanical (disturbed blood flow and high vessel wall stress), metabolic (hyperlipidemia, hyperglycemia), and inflammatory and thrombotic factors (3, 4). Plaque formation appears to be driven by altered endothelial cell (EC) phenotype, including increased permeability that allows lipoproteins to enter the vessel wall, increased expression of proteins that recruit monocytes, and other immune cells; activated ECs also stimulate smooth muscle cell migration and growth to form a neointima that narrows the vessel lumen. Further, vascular wall inflammation increases sensitivity of ECs to transforming growth factor–β (TGF-β), leading to endothelial-mesenchymal transition (EndMT) (5). The resultant EC-derived mesenchymal cells drive further increases in permeability and inflammation, thus perpetuating the incomplete repair state.

A similar process occurs in vascular malformations such as cerebral cavernous malformation (CCM) and hereditary hemorrhagic telangiectasia (HHT) (6, 7). In the inherited form of these diseases, patients have a heterozygous loss-of-function mutation in a key gene (CCM1, 2, or 3 for CCM; Alk1, endoglin, or Smad4 for HHT). A “second hit” mutation plus angiogenic or inflammatory stimuli lead to clonal expansion and formation of lesions with poorly formed vessel walls that are prone to rupture and bleeding. Again, the pattern of elevated permeability, abnormal extracellular matrix (ECM), inflammation, and EndMT can form positive feedback circuits that contribute to lesion progression. Yet another example is chronic inflammation due to persistent infection, injury, or autoimmune processes. In these conditions, angiogenic blood vessels remain immature with increased permeability, ECM remodeling, and expression of genes for leukocyte recruitment, thus perpetuating inflammation (8).

There are several well-conserved features that distinguish stable from remodeling vasculature. In stable vessels, ECs rest on a basement membrane of mainly laminins and type IV collagen with associated glycoproteins but low amounts of provisional ECM proteins such as fibronectin and fibrin (4). By contrast, all forms of instability are associated with metalloprotease secretion, ECM degradation, assembly of provisional matrices, and alterations in key EC signaling and gene expression pathways. Stable vessels produce high amounts of angiopoietin 1, which binds its EC receptor Tie2 to stabilize the vessels; expression of angiopoietin 2 (Ang2), an antagonist of Tie2, is low (9). Junctional permeability is low in stable vessels and elevated during inflammation and remodeling (8). This rule has exceptions because fenestrated endothelium allows passage of fluids, and the high endothelial venules that facilitate leukocyte movement into tissues during immune surveillance allow passage of cells; however, these forms of permeability have distinct mechanisms. All forms of remodeling are also associated with a shift toward a more oxidative state, or in the case of pathologies, to oxidative stress (10).

Persistent instability and disease

Local signals activate blood vessel endothelial cells (ECs), triggering events that transform ECs and create further instability. This is a feedforward cycle, as instability leads to matrix remodeling, leukocyte recruitment, inflammation, and EC transformation, which accelerates matrix remodeling and so on. Long duration of this process may lead to numerous pathologies. ROS, reactive oxygen species.


Whereas stable vessels exhibit normal EC fate markers, vascular instability is associated with EndMT, in which endothelial markers are reduced and mesenchymal markers increase (11). Functional consequences include increased permeability, increased expression of leukocyte adhesion molecules, enhanced motility, and expression and deposition of provisional ECM. EndMT has been observed in physiological and pathological angiogenesis, and in atherosclerosis, chronic inflammation, and vascular malformations.

How do physiological and pathological remodeling differ? In physiological remodeling, the initiating stimulus disappears once adaptation is completed, leading to restoration of normalcy. By contrast, in pathological remodeling, the initial triggers generally persist. Even when initiating triggers are removed, the vasculature does not return to its normal state. Such persistence involves positive feedback loops that maintain the endothelium in the remodeling or disease state, leading to disease progression. Several examples illustrate this point.

In atherosclerosis, statins slow progression but do not stop or reverse the disease even when plasma lipids are drastically lowered; ECs overlying atherosclerotic plaques remain activated. Some positive feedback loops involve increased vascular permeability and expression of leukocyte adhesion molecules. Entry of circulating ECM proteins such as fibronectin and fibrinogen into the vessel wall contributes to deposition of proinflammatory provisional ECM. Leaky endothelium also allows entry of lipoproteins into inflamed vessel wall under oxidative stress, resulting in production of proinflammatory oxidized lipoproteins that further increase lipid uptake and inflammatory mediators. Activation of adaptive immune responses to damaged proteins within the plaque and enhanced entrance of circulating inflammatory cells due to increased expression of adhesion molecules accelerate disease (3). Similar processes take place in transplant arteriopathy, characterized by low-grade, persistent vascular inflammation and incomplete repair (12). CCMs also exhibit this constellation of increased permeability, altered ECM, inflammatory activation, and EndMT that likely mediate disease progression (7).

These findings imply that persistent instability is central to multiple vascular disorders (see the figure). EndMT is likely an important component due to conversion of ECs to migratory mesenchymal cells that cause irreversible changes in tissue architecture. EndMT is induced by inflammation-driven decreases in protective fibroblast growth factor (FGF) signaling in ECs, which increases sensitivity to TGF-β (11). Thus transformed, ECs increase expression of leukocyte adhesion molecules and altered ECM, and have high vascular permeability, further driving inflammation. These events establish a positive feedback loop in which EndMT drives inflammation, which leads to further EndMT. This inflammation–TGF-β circuit also promotes fibrosis, which is largely irreversible.

This hypothesis implies that interventions to restore vascular stability might improve outcomes in multiple disease states. This might be achieved by agonists for Tie2, FGF receptors, or other pathways that promote stability, or by antagonists for inflammatory or endothelial TGF-β pathways that promote instability. The key requirement is to break the feedback loops that maintain the remodeling state and thereby restore tissue homeostasis. Although this notion flies in the face of current trends toward personalized medicine, the search for unifying principles that govern health and disease is as old as medicine. Much remains to be done before we achieve the deep understanding of biological processes needed to design effective interventions in vascular disease, yet our current state of knowledge offers a glimpse of the path forward.


View Abstract

Navigate This Article