Relevant Growth Factors
Vascular endothelial growth factor (VEGF) is the central growth factor in angiogenesis. It is not only mitogenic for endothelial cells, but is proangiogenic by influencing effectors of other important steps in the vascular remodeling process, including the following: (a) cell signaling (nitric oxide (NO) production); (b) remodeling of the extracellular matrix (upregulation of urokinase- and tissue-type plasminogen activator (uPA, tPA), PA inhibitor (PAI-1), uPA receptor (uPAR), and matrix metalloproteinase (MMP)); and (c) promoting chemotaxis to assist productive migration of cells in tube formation. Furthermore, VEGF is required to maintain vascular integrity because in its absence there can be an attrition (rarefaction) of vessels in the tissue.
Deletion of the VEGF gene, or eliminating its action by gene deletion of one of its receptors, is lethal to embryonic development (Table 1). VEGF is encoded by a single gene that is posttranscriptionally spliced into several different isoforms (isoforms 121, 145, 165, 189, and 206). The most prevalent isoform is VEGF165. The gene contains a signal-sequence characteristic of protein designed for export from the cell. VEGF is secreted by numerous tissues, including endothelial, smooth muscle, and skeletal muscle cells. All but one isoform of VEGF possess a heparin-binding domain, thereby facilitating binding to the heparan sulfates in the extracellular matrix (ECM). This establishes an extracellular pool of VEGF that is available for action upon degradation of the ECM. The VEGF gene contains an upstream regulatory sequence that increases VEGF mRNA production when bound by the hypoxia inducible factor (HIF1). This is the same sequence and regulatory factor that enhances erythropoietin gene expression during hypoxia. Control of the HIF1 protein abundance appears to be a combination of its ubiquitination state and gene transcription. Whereas hypoxia is an important stimulus influencing HIF-regulated VEGF expression, stimuli other than hypoxia (e.g., NO) can act through the HIF regulatory process.
VEGF has two high-affinity receptors: flt-1 and flk-1/KDR. Both are tyrosine kinase receptors found exclusively on endothelial cells. Flt-1 is an important binding site for VEGF-mediated vasculogenesis during development. In adult tissue, though, critical aspects of VEGF induced angiogenesis (activation, proliferation, migration) can be mediated by binding the KDR receptor; thus flt-1 is thought to play a minor functional role in angiogenesis after development.
The angiopoietins (Ang1 and Ang2) are recently discovered effectors of vascular development and remodeling. They are essential for normal embryonic development, as lethality occurs with gene deletion (Table 1). Ang1 promotes maturation and stabilization of vessels through binding to and activation of the endothelial cell-specific tyrosine kinase receptor, Tie-2. In contrast, Ang2 augments angiogenesis via binding to, but not activating, Tie-2. The ability of Ang2 to augment angiogenesis is, in part, the result of preventing Ang1 from binding to and activating Tie-2. Therefore, Ang2 displaces Ang1 and destabilizes the vasculature and makes it more responsive to VEGF. Thus, Ang1 and Ang2 are natural competitors in regulating vascular remodeling. Ang1 is constitutively expressed throughout the body, whereas Ang2 is apparent only at vascular sites active in vessel growth or remodeling. Ang1 and Ang2 are not mitogenic to endothelial cells; rather, they are dependent on VEGF. Thus interactions between VEGF and angiopoietin regulation are likely critical in exercise-induced angiogenesis.
Other growth factors
Fibroblast growth factors are a family of angiogenic growth factors that are mitogenic to all three cell types that comprise the vasculature: endothelial cells, smooth muscle cells, and fibroblasts. Probably the best-studied protein is FGF-2 (also referred to as basic FGF). Deletion of the FGF-2 gene is not lethal, as the animals are healthy and appear to grow normally, although some nonvascular deficits have been described. Thus, FGF-2 is not obligatory for normal development of the vasculature. Although its role in some aspects of vascular remodeling (e.g., wound healing) is recognized, its importance in angiogenesis has not received the attention of VEGF. Similarly, although TGFb is implicated in some aspects of angiogenesis, it may work through VEGF and has not merited the attention as a primary initiator of angiogenesis. Other growth factors associated with angiogenesis include PDGF, TNFa, and epidermal growth factor (EGF) (for review, see (8)).
The Angiogenic Process
Angiogenesis can occur by two separate mechanisms: sprouting and intussusception. Intussusception refers to the formation of pillar-like structures within an existing capillary lumen (Fig. 3). Formation of the pillar divides the capillary into two luminal regions, ultimately yielding the formation of two distinct capillaries. In contrast to intussusception, capillary sprouting requires degradation of the basement membrane and extravasation of the activated endothelial cells into the extravascular space (Fig. 3). The mechanisms that distinguish whether a capillary undergoes intussusception or sprouting are not well understood. At the present time, most of our information is related to sprouting angiogenesis and this will be the focus of our discussion.
As illustrated in Figure 4, there are several critical steps thought to occur in angiogenesis. Interruption at any step can suspend or eliminate the angiogenesis process. First, there is endothelial cell activation, as occurs with binding of VEGF to its unique receptor flk-1/KDR on the endothelial cell. Activation may also be initiated by shear stress, by way of binding sites for shear stress response elements on gene promoters, or by indirect upregulation of gene expression by NO as is the case for VEGF. This raises the potential for vascular remodeling caused by hemodynamic conditions in the circulation. Upon activation, the endothelial cells begin to proliferate and, depending upon a coordination of events, the angiogenic process continues. Second, degradation of the basement membrane and extracellular matrix, which anchor the capillary in place, must proceed. Degradation is initiated by the action of the plasminogen activators (uPA, tPA). As their names imply, these proteins convert inactive plasminogen to active plasmin, which in turn converts the inactive, pro-matrix metalloproteinases (proMMPs) into active MMPs. In addition to plasmin, membrane-type MMPs (MT-MMPs) also convert proMMPs to active MMPs. It is the active MMPs that are primarily responsible for the degradation of the basement membrane and ECM. Once the basement membrane and ECM have been degraded, activated endothelial cells can migrate and continue the angiogenic process. Third, to initiate tube formation and progression, migrating endothelial cells are guided by surface adhesion molecules and integrins (i.e., avb3). Positioning of the integrins and MT-MMPs, as well as release of matrix-bound growth factors, is localized to the leading edge of new capillary formation. This helps regulate capillary tube growth and/or minimize unnecessary and uncontrolled capillary growth. To be a functional capillary, the newly formed tube must rejoin an existing capillary that is in the vascular bed. This requires coordination of chemotaxic signals between the recipient site and the developing capillary tube. The newly forming sprout is thought to be guided by several mechanisms: differential expression and density of eph and ephrin proteins on endothelial cells and surrounding tissue, which likely alter integrin expression and signaling, particularly of avb3, thereby either enhancing or inhibiting the anchoring of newly formed endothelial cells; enhanced chemotaxis of endothelial cells when VEGF simultaneously binds both KDR and neuropilin-1 receptors; and possibly by tensional forces exerted through the extracellular matrix by preexisting capillaries. These mechanisms lead to well placed but structurally immature capillaries that are excessively permeable and highly leaky to plasma. Fourth, the final critical step in angiogenesis requires stabilization and maturation of the newly produced capillary tube. This is mediated, in part, by changes in expression of angiopoietins and integrins. Increased Ang1 and reduced Ang2 dominance assist in pericyte recruitment to surround the endothelial cord to form a rigid structure for mature tube formation. This completes maturation to a nonleaky, functioning capillary indistinguishable from the originating capillary.
Whereas the development of an enhanced capillarity within muscle is predicated upon the development of unique cellular and molecular events to remodel the tissue, it is important to recognize operational features about the exercise bout that could influence the appearance and/or magnitude of these cellular and molecular signals. First, it is obvious that the motor units within a region of muscle must be recruited in order for an angiogenic stimulus to be developed, because the angiogenic adaptation to training is not a systemic, whole-body response. Thus the type, intensity, and duration of exercise are important. Further, altered recruitment of motor units may occur as an adaptation to training. Thus the developed angiogenic stimuli could vary over time of training, even when performing the same daily exercise task. This has significant implications in determining the magnitude of the final angiogenic adaptation and/or the progression in daily exercise intensity/duration, as the training program advances, that may be needed to optimize the final adaptive response. Second, as alluded to above, the inherent vascularity of the tissue, which could modify important hemodynamic and/or metabolic signals, could significantly temper the development of an angiogenic stimulus, even for the same given recruitment of motor units. Thus the adaptive responses of the low capillary density, fast-twitch white region could be quite different from the high capillary density, fast-twitch red region of the same muscle. Third, the formation of the enhanced capillary network will take time to develop. Whereas important signals are expected to be produced during and following an individual exercise bout, their cumulative effect in tissue remodeling is only evident over time. Furthermore, retention of the enhanced capillarity, developed by chronic exercise, may not occur following cessation of the stimuli that stimulated the angiogenesis in the first place. Although there is evidence to suggest that mature capillaries are relatively stable and retained, inactivity could prove costly in loosing this adaptation. Finally, as developed above, the process of angiogenesis is sufficiently complex, requiring an integration of a multitude of differently controlled cellular and tissue events, which appropriate exercise stimuli must coordinate, sustain, and/or possibly increase, to achieve the final outcome of an enhanced capillarity.
Unfortunately, we have little definitive evidence to indicate what transducing stimuli, prompted by muscle contraction, lead to the cellular and molecular events producing the increase in capillarity (5). Transducing stimuli that have been identified, or may be involved, include 1) hemodynamic stimuli, 2) metabolic stimuli, 3) tension developed during muscle contraction, and 4) factors extrinsic to the muscle. First, hemodynamic-derived stimuli are caused by changes in wall tension, defined as the radial force within the vessel wall needed to withstand the internal blood pressure, and/or shear stress, defined as the drag or pull experienced by endothelial cells by the flow of blood across their surface. Significant changes in wall tension and shear stress can, in turn, prompt activation of the angiogenic process and lead to vascular remodeling. For example, a continuous increase of blood flow to muscle, established experimentally, can elevate muscle capillarity. Further, maintaining the size and wall thickness of large conduit vessels requires the normal high internal pressure and flow characteristic of these arterial vessels. Interestingly, removal of the endothelium eliminates this remodeling caused by changes in hemodynamic stimuli. Thus the endothelium is an indispensable component in vascular signaling and remodeling. Second, the best characterized metabolic stimulus that is likely important in prompting angiogenic signals is the challenge of a low po2. The cell “senses” a low po2 and, as described above, activates the HIF transcription factor. This upregulates the production of VEGF mRNA and subsequent production of this important angiogenic cytokine. Third, force developed during muscle contractions could be an important contributor to angiogenesis. Distending force has been shown to be important in the migration and organization of developing capillaries. This could account for disparate observations in the control of angiogenesis during muscle contractions, with and without a weight-bearing load (6,7). Finally, recent evidence raises the possibility that hematopoietic-derived endothelial-precursor cells contribute to the endothelial cell proliferation associated with vascular remodeling. This of course requires that some yet undescribed signal provides the attraction directing these blood-borne cells to locate at the site of vascular remodeling. Whether these processes account for the exercise-induced vascular remodeling remains to be determined. However, important new information has become available.
Control of Angiogenesis by Exercise
Exercise leads to a marked upregulation of VEGF mRNA within the active muscle (2,13). The increase is greatest (two- to eightfold) shortly after the exercise bout (2–4 h) and declines with time thereafter, returning to normal within 24 h. There is a coordinated increase in VEGF protein that accumulates between the fibers in the intracellular matrix, implying a rapid translation of the newly produced mRNA. Thus the involvement of this critical cytokine is implicated in the exercise-induced angiogenesis.
The upregulation of VEGF mRNA is dependent upon the intensity of exercise, tempered with subsequent days of exercise at the same exercise effort, and modest to absent in the well-trained state. Because hypoxia can regulate VEGF transcription, low tissue po2 is common in working muscle, and corresponding increases in HIF are observed in active muscle (3), it is credulous to expect that low po2 is the critical cellular signal prompting the increase in VEGF mRNA. However, the role for hypoxia, as the sole determinant in the upregulation of VEGF mRNA during normal exercise is not clear. The increase in VEGF mRNA is considerable in well-perfused muscle during moderate-intensity exercise and not appreciably, if at all, exaggerated with the same exercise during hypoxia (9). Whereas low po2 is competent to upregulate VEGF mRNA, it is likely that factors other than low po2 also contribute to the upregulation of VEGF in muscle after activity.
That hemodynamic stimuli, related to wall tension and/or shear stress, contribute to the angiogenesis induced by exercise remains a possibility. There is evidence that elevated blood flow to muscle, in the absence of muscle contractions, does not prompt VEGF mRNA upregulation (11). This absence of a VEGF response is striking and argues that, unlike the hemodynamic conditions in larger conduit vessels, capillary remodeling may be unrelated to wall tension and/or shear stress; however, there is no assurance that the flow dynamics created by elevating flow experimentally simulate the flow dynamics present in contracting muscle caused by mechanical forces and pulsatile flow gradients. Furthermore, other evidence demonstrates that muscle capillarity is increased when blood flow is chronically elevated experimentally. Thus hemodynamic stimuli remain as potentially important signals initiating angiogenesis induced with exercise training.
There is an increase in both flk/KDR and flt-1 mRNA in active muscle with exercise (1). This upregulation implies an increase in the ability to respond to the increase in VEGF signal, a response that could prompt capillary growth. However, although the presence of VEGF appears essential to initiate and facilitate angiogenesis, its actions are not sufficient as the sole agent to complete the process. As described above, the angiopoietins are essential for normal vascular remodeling. Preliminary evidence indicates that exercise, particularly ischemic exercise, leads to a shift in Ang1/Ang2 mRNA expression ratio that would support destabilization of the vasculature. How this response is controlled and coordinated to the response in VEGF regulation remains to be clarified.
The importance of FGF-2 in the exercise-induced muscle angiogenesis is questioned, because changes in mRNA abundance are modest to absent (3) and there is no change in FGF-2 protein content in the active muscle. Similar questions can be raised about the role of TGFb. There is an absent to modest increase in TGFb mRNA following exercise. Furthermore, for reasons described above, the roles of FGF-2 and TGFb in angiogenesis are viewed as not as important as that of VEGF.
There is evidence that normal NO production is an important link in the signaling for angiogenesis induced by muscle contractions. Inhibition of normal NO production eliminated the increased vascularity normally found when muscle contractions were introduced experimentally by chronic electrical stimulation (6), but not when introduced with weight bearing during treadmill running (7). If developed muscle force is appreciably different between these treatments, it suggests that tissue tension is critical to the process of angiogenesis with active muscle. How this may be critical to some step in the angiogenic process is not known. Additional insight into the control of angiogenesis comes from experimentally controlling the dissolution of the extracellular matrix by the actions of MMP. Inhibition of MMP activity during the period of muscle contractions eliminated the normal elevation in capillarity within the muscle (4). Endothelial cell activation and proliferation appears to be evident, but migration was blocked without dissolution of the basement membrane and extracellular matrix. These results illustrate the complex and intriguing steps that control angiogenesis. Future research will likely provide striking answers to this fascinating process.
Even though blood flow capacity of skeletal muscle is extremely high, well in excess of that typically used even during maximal whole-body exercise, exercise training can increase it further. This is most easily observed in the relatively low-flow regions of muscle and can be attributed to two causes: structural enlargement of conduit vessels to increase flow potential, and endothelial-mediated dilatory adaptations that probably ensure optimal use of the vessel caliber. Unfortunately, there are few studies evaluating changes in large vessel diameter with exercise training. Furthermore, it is likely that the high inherent flow capacity and the inadequate nature of any training stimulus, caused by relatively modest durations of experimentally increased flow demands, would undermine effective studies. However, there is compelling evidence that high flow demands through conduit arteries remodel vessels to become larger thicker-walled vessels. This is nicely illustrated by the work of Tronc et al. (Fig. 5), where blood flow through the common carotid was elevated surgically using an arteriovenous (a-v) shunt (12). The high-flow region of the carotid that was proximal to the a-v shunt greatly increased its diameter, which in turn returned shear stress to normal, whereas the low-flow region that was distal to the a-v shunt exhibited a reduction in vessel diameter. Interestingly, this response induced by an exceptionally high blood flow was eliminated when normal NO production was disrupted. In the context of exercise, however, vascular remodeling of conduit vessels is most realistically evaluated in experimental models where higher blood flow during exercise is directed through relatively unused conduit vessels. This is easily achieved in animal models of peripheral arterial insufficiency where flow capacity to the calf muscles is experimentally limited by occlusion of the femoral artery. The reduced distal pressure alters flow patterns to direct flow through a few preexisting conduits that serve as collateral vessels, which circumvent blood around the obstruction. Exogenous administration of FGF-2 or VEGF increases the diameter of these vessels and markedly improves blood flow downstream. Similarly, an appreciable increase in blood flow to the calf muscles occurs when the animals are subjected to treadmill walking, owing to a significant increase in caliber of these vessels. This model provides an opportunity to evaluate the process of arteriogenesis induced by exercise. It also gives insight into the process by which enhanced physical activity can prove so valuable in managing patients with intermittent claudication.
The prompting stimulus for these conduit vessels to remodel is likely local in nature, specific to the vessels affected by occlusion of the femoral artery. Unlike the situation in the distal muscle at risk of ischemia, the involvement of hypoxia as a prompting stimulus is unlikely, since these arteries are perfused with high-po2 blood derived from the aorta. The affected endothelium should experience a similar high po2, far distant from that shown to upregulate VEGF and prompt angiogenesis. On the other hand, hemodynamic factors related to the altered flow and pressure pattern remain implicated. Occlusion of the primary arterial supply dramatically alters the pressure gradient within the vascular tree and, in time, increases flow through the vessel in an attempt to support flow demands downstream. The decrease in transmural pressure that occurs in the collateral vessel is opposite that typically needed to increase vessel caliber; thus, attention can be placed on an increased shear stress caused by the increased blood flow through the vessel as the important hemodynamic stimulus initiating arteriogenesis. This places emphasis on the function of the endothelium in the transduction of signals involved in vessel enlargement. Experimental evidence demonstrates that this emphasis is well placed. For example, vessel enlargement and increased collateral blood flow prompted by exogenous delivery of recombinant angiogenic growth factors occur only in the presence of vascular occlusion and not in the contralateral limb in the absence of occlusion, even though an effective dose was delivered to both. This implies that the altered hemodynamic conditions promulgated by occlusion of the primary artery lead to cellular changes that now make it responsive to the circulating angiogenic growth factor. This might reasonably occur if enhanced shear stress upregulates growth factor receptors in the endothelial cell. Enhanced binding of growth factor could activate the endothelial cells to initiate vascular remodeling. Experimental evidence is needed to support this simple hypothesis. As illustrated in Figure 6, the potential importance of shear stress is also evident by the additive effect on collateral blood flow by the combination of treatments, exogenous growth factor administration and exercise training (14). This increase would be expected if the added shear stress, caused by the flow increases during exercise, served to enhance growth factor efficacy in arteriogenesis. The importance of the endothelial cell is further implicated by altering the endothelial-mediated NO signaling pathway (15). Inhibition of NOS with L-NAME eliminates vascular remodeling and the increase in collateral blood flow induced by exercise training (7). Thus, a normal NOS function is essential for arteriogenesis to develop. As discussed above, angiogenesis in the active muscle of these same trained animals was not eliminated. Muscle capillarity increased, identical to animals similarly trained but without any inhibition of NO production. This illustrates that fundamental differences exist between the control of angiogenesis and arteriogenesis. As we learn more about the process of exercise-induced vascular remodeling that increases blood flow capacity, we are likely to better understand some health related benefits of physical activity.
Exercise imparts a powerful stimulus for vascular remodeling evident by an increase in capillarity within the active muscle (angiogenesis) and an enlargement of conduit vessels (arteriogenesis) increasing flow capacity to muscle, especially when compromised by vascular obstruction. The vascular remodeling is intricate and involves a complex coordination among angiogenic growth factors (e.g., VEGF), receptors, and modulating influences including the angiopoietins and ephrins. In the case of angiogenesis, there must be an integrated and coordinated sequence of events that include endothelial cell activation, proliferation, and migration; basement membrane and extracellular matrix dissolution; tube formation; and attachment to the recipient capillary with maturation to stabilize the capillary by pericyte envelopment and matrix reformation. Arteriogenesis requires remodeling of all cell types that constitute the vascular system (endothelial, smooth muscle, and fibroblast cells) to effect enlargement of the larger supply vessels. These vascular adaptations serve to enhance muscle performance by increasing the muscle’s oxygen-exchange capacity and by increasing blood flow capacity, especially when limited by upstream vascular obstruction. These adaptations potentially contribute to the significant cardiovascular performance and health benefits realized by individuals who are physically active.
The authors thank Don Connor and Howard Wilson for their assistance with the illustrations. This work and the cited work of the authors was supported by NIH grants HL37387, HL10406, and HL10485.
1. Gavin, T. P., and P. D. Wagner. Attenuation of the exercise-induced increase in skeletal muscle Flt-1 mRNA by nitric oxide synthase inhibition. Acta Physiol. Scand. 175: 201–209, 2002.
2. Gustafsson, T., and W. E. Kraus. Exercise-induced angiogenesis
-related growth and transcription factors in skeletal muscle, and their modification in muscle pathology. Frontiers Biosci. 6: D75–D89, 2001.
3. Gustafsson, T., A. Puntschart, L. Kaijser, E. Jansson, and C. J. Sundberg. Exercise-induced expression of angiogenesis
-related transcription and growth factors in human skeletal muscle. Am. J. Physiol. 276: H679–H685, 1999.
4. Haas, T. L., M. Milkiewicz, S. J. Davis, A. Zhou, S. Egginton, M. D. Brown, J. A. Madri, and O. Hudlicka. Matrix metalloproteinase activity is required for activity-induced angiogenesis
in rat skeletal muscle. Am. J. Physiol. 276: H679–H685, 1999.
5. Hudlicka, O., M. Brown, and S. Egginton. Angiogenesis
in skeletal and cardiac muscle. Physiol. Rev. 72: 369–417, 1992.
6. Hudlicka, O., M. D. Brown, and H. Silgram. Inhibition of capillary growth inchronically stimulated rat muscles by N(G)-nitro-1-arginine, nitric oxide synthase inhibitor. Microvasc. Res. 59: 45–51, 2000.
7. Lloyd, P. G., H. T. Yang, and R. L. Terjung. Arteriogenesis
in rat ischemic hindlimb: role of nitric oxide. Am. J. Physiol. 281: H2528–H2538, 2001.
8. Papetti, M., and I. M. Herman. Mechanisms of normal and tumor-derived angiogenesis
. Am. J. Physiol. 282: C947–C970, 2002.
9. Richardson, R. S., H.Wagner Wagner, S. R. Mudaliar, R. Henry, E. A. Noyszewski, and P. D. Wagner. Human VEGF
gene expression in skeletal muscle: effect of acute normoxic and hypoxic exercise. Am. J. Physiol. 277: H2247–H2252, 1999.
10. Robinson, D. M., R. W. Ogilvie, P. C. Tullson, and R. L. Terjung. Increased peak oxygen consumption of trained muscle requires increased electron flux capacity. J. Appl. Physiol. 77: 1941–1952, 1994.
11. Roca, J., T. P. Gavin, M. Jordan, N. Siafakas, H. Wagner, H. Benoit, E. Breen, and P. D. Wagner.. Angiogenic growth factor mRNA responses to passive and contraction-induced hyperperfusion in skeletal muscle. J. Appl. Physiol. 85: 1142–1149, 1998.
12. Tronc, F., M. Wassef, B. Esposito, D. Henrion, S. Glagov, and A. Tedgui. Role of NO in flow-induced remodeling of the rabbit common carotid artery. Arterioscler. Thromb. Vasc. Biol. 16: 1256–1262, 1996.
13. Wagner, P. D. Skeletal muscle angiogenesis
. A possible role for hypoxia. Adv. Exptl. Med. Biol. 502: 21–38, 2001.
14. Yang, H. T., R. W. Ogilvie, and R. L. Terjung. Exercise training enhances basic fibroblast growth factor-induced collateral blood flow. Am. J. Physiol. 274: H2053–H2061, 1998.
15. Yang, H. T., Z. Yan, J. A. Abraham, and R. Lg Terjun. VEGF121 and bFGF-induced increase in collateral blood flow requires normal nitric oxide production. Am. J. Physiol. 280: H1097–H1104, 2001.
Keywords:©2003 The American College of Sports Medicine
angiogenesis; arteriogenesis; VEGF; angiopoietin; extracellular matrix