Impaired Autonomic Regulation of Resistance Arteries in Mice With Low Vascular Endothelial Growth Factor or Upon Vascular Endothelial Growth Factor Trap Delivery
Background— Control of peripheral resistance arteries by autonomic nerves is essential for the regulation of blood flow. The signals responsible for the maintenance of vascular neuroeffector mechanisms in the adult, however, remain largely unknown.
Methods and Results— Here, we report that VEGF∂/∂ mice with low vascular endothelial growth factor (VEGF) levels suffer defects in the regulation of resistance arteries. These defects are due to dysfunction and structural remodeling of the neuroeffector junction, the equivalent of a synapse between autonomic nerve endings and vascular smooth muscle cells, and to an impaired contractile smooth muscle cell phenotype. Notably, short-term delivery of a VEGF inhibitor to healthy mice also resulted in functional and structural defects of neuroeffector junctions.
Conclusions— These findings uncover a novel role for VEGF in the maintenance of arterial neuroeffector function and may help us better understand how VEGF inhibitors cause vascular regulation defects in cancer patients.
Received December 4, 2009; accepted May 24, 2010.
Several cues have been identified that control the development of the sympathetic nervous system,1 but the molecular mechanisms underlying the maintenance of autonomic innervation of peripheral resistance arteries in adulthood remain largely unknown. Periarterial nerves do not penetrate the medial vascular smooth muscle cell (SMC) layer but are confined to the media adventitia border. The varicose terminal portions of these nerve fibers release transmitter en passage,2 and autonomic neuroeffector junctions (NEJs) in resistance arteries are “atypical” in that they contain prejunctional membrane thickenings with synaptic vesicles but lack specialized postjunctional structures.2
Clinical Perspective on p 281
By releasing neurotransmitters, autonomic nerves control the blood supply to organs. For instance, autonomic control of resistance arteries mediates thermoregulation in the skin and redistribution of flow from internal organs to the brain in stress conditions.3,4 Dysfunction of the periarterial autonomic nervous control can lead to hypertension, Raynaud phenomenon, and orthostatic hypotension.5 Impaired vascular regulation has also been implicated in the hand-foot syndrome in cancer patients receiving vascular endothelial growth factor (VEGF) inhibitors.6 This syndrome is characterized by dysesthesia, erythema, and tingling of extremities, which may progress to burning pain with dryness, cracking, and ulceration of the skin.7,8 The molecular basis of this syndrome remains largely enigmatic.
VEGF, a key regulator of angiogenesis in health and disease,9 also has neurotrophic effects. VEGF stimulates axon outgrowth and survival of neurons in vitro,10 and low VEGF levels in VEGF∂/∂ mice, engineered to lack the hypoxia-response element in the VEGF promoter, cause adult-onset motoneuron degeneration, reminiscent of the human motoneuron degenerative disorder amyotrophic lateral sclerosis.11 However, surprisingly little is known about the effects of VEGF on peripheral autonomic nerves and, in particular, about its possible role in the autonomic innervation of resistance arteries in vivo. In vitro studies show that VEGF stimulates axon outgrowth and survival of superior cervical ganglia12,13 and induces growth cone spreading of cultured postganglionic neurons.14 Furthermore, VEGF promotes the reinnervation of denervated femoral arteries in vivo.14 However, a role for VEGF in the maintenance of the perivascular autonomic nerve plexus in vivo has not been addressed. Here, we assessed whether VEGF might modulate the autonomic nervous control of resistance arteries.
Transgenic Mice and Soluble Flk1 Gene Transfer
VEGF∂/∂, VEGF-LacZ, and Flt1-LacZ mice were described previously.11,15,16 Expression of the soluble extracellular domain of Flk1 (sFlk1) in the circulation was obtained by hydroporating mice with 2.5 mL Ringer solution (150 mmol/L NaCl, 4 mmol/L KCl, 1 mmol/L CaCl2) containing 50 μg per mouse of a vector encoding sFlk1 or an empty vector as control.17 After 5 days, plasma protein levels of sFlk1 were determined with a soluble VEGF receptor-2 (VEGFR-2)–specific immunoassay (R&D Systems, Minneapolis, Minn).
Histology and Immunohistochemistry
Perfusion-fixed arteries were embedded in paraffin, and sections were stained with the following antibodies: polyclonal rabbit anti-smoothelin antibody (1:100; a gift from G. van Eys, Maastricht University), anti-desmin antibody (1:50; Cappel, Durham, NC), and anti–SM myosin heavy chain (MHC) antibody (1:400; ab683, Abcam, Cambridge, UK). Procedures used for whole-mount staining of arteries, analysis of skin vascular density, LacZ enzymatic staining, alkaline phosphatase staining, glyoxylic acid staining, and noradrenaline content determination are described in the Methods section of the online-only Data Supplement.
Isometric Wire Myography
Isometric wire myography was performed as described previously.18 Detailed procedures can be found in the Methods section of the online-only Data Supplement.
Quantitative Real-Time Polymerase Chain Reaction Experiments
Messenger RNA transcript levels were quantified and normalized relative to the expression level of β-actin or GAPDH with the use of premade primers and probes (TaqMan Gene Expression Assays, Applied Biosystems, Foster City, Calif). Relative gene expression in VEGF∂/∂ versus control arteries was calculated with the 2−ΔΔCT method.
Transmission Electron Microscopy
Standard procedures were used for transmission electron microscopy. Detailed procedures can be found in the Methods section of the online-only Data Supplement.
A rectal probe connected with a digital thermometer (Physitemp Instruments) was used to measure body temperature in an ambient temperature of 21°C on 3 consecutive days with 3 measurements per day; the average temperature was recorded as baseline. During exposure to cold stress (8°C for VEGF∂/∂ mice, 4°C for sFlk1 mice), rectal temperature was measured at half-hour intervals for 4.5 hours.
Blood Flow Measurements
Details on these and other methods used in this study can be found in Methods section of the online-only Data Supplement.
For comparison of 2 genotypes or treatment groups, the unpaired Student t test was used. When data were not normally distributed, the Mann–Whitney test was used; medians and interquartile ranges are reported. For the evaluation of body temperature evolution during cold stress experiments, repeated-measures ANOVA was used. For statistical analysis of concentration-response curves, ANOVA was used with Bonferroni correction for multiple testing. A limitation of our study is that, because of the limited availability of transgenic mice, sample size in our experiments is often small, increasing the risk for a type I error.
Expression of VEGF and VEGFRs
For VEGF to regulate the function of resistance arteries, VEGF and VEGFRs should be expressed in these arteries. We therefore determined the spatial expression pattern of VEGF and VEGFRs (VEGFR-1/Flt1 and VEGFR-2/Flk1) in mesenteric and saphenous arteries. In VEGF-LacZ mice, which express nuclear LacZ under the control of the endogenous VEGF promotor,15 X-gal staining revealed VEGF expression by medial SMCs (Figure 1A) and by neurons in celiac and superior cervical ganglia (Figure 1B and not shown), implying that VEGF may also be released from nerve terminals. X-gal staining of arterial cross sections of Flt1-LacZ mice, which express cytosolic LacZ under the control of the endogenous Flt1 promotor,16 revealed Flt1 expression in medial SMCs (Figure 1C). This expression pattern of Flt1 was confirmed by a VEGF-B186-alkaline phosphatase (AP) fusion protein, which selectively binds to Flt1 (Figure 1D). Staining of whole-mount arteries with a VEGF-AP fusion protein, which binds to Flk1, Flt1, and neuropilin, revealed expression of VEGFRs also on periarterial nerves (Figure 1E), and immunostaining for Flt1 or Flk1 labeled perivascular nerve fibers and endings (Figure 1F and 1G). Thus, VEGF and VEGFRs are expressed in resistance arteries and periarterial nerves. We therefore tested whether VEGF might play a role in vascular regulation of peripheral resistance arteries.
To explore whether VEGF affected vascular regulation, we used VEGF∂/∂ mice, which express reduced VEGF levels.11 ELISA confirmed that VEGF protein levels were reduced in VEGF∂/∂ arteries and ganglia (note I in the online-only Data Supplement), and immunostaining and quantitative polymerase chain reaction revealed that VEGF expression was reduced in arterial SM and postganglionic neurons (Figure I and note I of the online-only Data Supplement). Because VEGF∂/∂ mice develop motoneuron degeneration beyond 5 to 7 months of age,11 we used only presymptomatic VEGF∂/∂ mice (≤3 months of age) for all experiments.
Impaired Vasoconstrictor Responses of VEGF∂/∂ Resistance Arteries Ex Vivo
We then explored by isometric wire myography ex vivo whether the reduced VEGF levels in VEGF∂/∂ mice impaired vascular regulation of isolated small muscular arteries (mesenteric and saphenous arteries) and large elastic arteries (common carotid artery).18 To study whether VEGF modulates vasoregulatory responses of SMCs, we induced vasoconstriction by exposing arteries to stimuli that have direct effects on SMCs through distinct signal transduction pathways, ie, high potassium (K+), angiotensin II, noradrenaline, or the thromboxane A2 analog U46619. Because the contractile response of an artery depends on the intrinsic SMC properties but also on the thickness of the SMC layer (and because VEGF∂/∂ arteries were generally smaller; Table I of the online-only Data Supplement), we not only measured active wall tension (N/m), defined as the force generated by the artery divided by the arterial segment length, but also calculated active wall stress (N/m2), a parameter of intrinsic SMC function, by correcting active wall tension for media thickness.
Compared with wild-type (WT) mice, contractile responses of resistance arteries to high K+, angiotensin II, noradrenaline, or U46619 were reduced in VEGF∂/∂ mice, whereas the sensitivity to these stimuli was not modified (Figure 2A, 2C, and 2D, plus Table II and Figure I of the online-only Data Supplement). Both wall tension and wall stress were reduced by >50% in VEGF∂/∂ mice (Table II of the online-only Data Supplement). The fact that arterial contractility was impaired regardless of the type of vasoconstrictor stimulus indicates that SMCs were generally dysfunctional rather than having a selective defect in a particular signaling pathway. In addition, because the contractile responses were also impaired in the presence of the nitric oxide synthase inhibitor nitro-l-arginine, the defect was not attributable to abnormal production of nitric oxide (not shown). Notably, the contractile response of the elastic carotid arteries was normal (Table II and Figure IA and ID of the online-only Data Supplement and Figure 2B), indicating that vascular regulation by densely innervated muscular arteries, but not by sparsely innervated elastic arteries, is impaired in VEGF∂/∂ mice. Vascular SMC dysfunction was not accompanied by endothelial cell dysfunction because endothelium-dependent relaxation induced by acetylcholine was normal in resistance arteries of VEGF∂/∂ mice (Figure II of the online-only Data Supplement).
Dedifferentiation of Vascular SMCs in VEGF∂/∂ Resistance Arteries
To characterize SMC dysfunction in VEGF∂/∂ arteries in more detail, we measured by quantitative polymerase chain reaction the expression of “early” and “late” SMC differentiation markers, which are important for establishing SMC contractile properties in a developmentally sequential order.19 At 3 weeks of age, ie, when the autonomic innervation of resistance arteries was just established,20 expression of early (nonmuscle MHC, desmin, and calponin) and late (SM MHC and smoothelin-B) SMC differentiation markers was comparable in WT and VEGF∂/∂ saphenous arteries (Table 1). In contrast, by 3 months of age, expression of early SMC markers was normal, but expression of the late SMC markers was reduced in VEGF∂/∂ saphenous arteries (Table 1).
These results were confirmed by immunostaining (Figure III of the online-only Data Supplement). No differences were found in the expression of SMC differentiation markers in common carotid arteries (Table 1). The reduced expression of late SMC differentiation markers in VEGF∂/∂ arteries can explain their generally impaired contractile properties. This defect appeared to be specific for the contractile machinery because expression of postjunctional neurotransmitter receptors (α1-adrenoreceptors, purinergic receptor P2X1, and NPY1 receptor) and neurotrophic factors (nerve growth factor, artemin, and neurotrophin-3) was not significantly altered in VEGF∂/∂ arteries (not shown). Thus, SMCs in innervated resistance arteries differentiate normally during postnatal development but later dedifferentiate in VEGF∂/∂ mice.
Neuroeffector Dysfunction in VEGF∂/∂ Arteries
Because vascular regulation was impaired in densely innervated, but not in sparsely innervated, arteries in VEGF∂/∂ mice, suggesting a possible dysfunction of perivascular nerves, we studied neuroeffector function by ex vivo myography. In mesenteric arteries of WT mice, electric field stimulation (EFS) or administration of the indirectly acting sympathomimetic tyramine induced a strong vasoconstrictor response (Figure 3A and 3B). In contrast, the contractile response was severely impaired in VEGF∂/∂ arteries (Figure 3A and 3B). To exclude that the impaired response to these neurogenic stimuli in VEGF∂/∂ mice reflected only intrinsic SMC (and not neuronal) defects, we expressed the responses as a fraction of the maximal response to exogenously applied noradrenaline. Even after normalization, the contractile response remained lower in VEGF∂/∂ mice (Figure 3C and 3D and Table 2), indicating that sympathetic neuroeffector regulation was dysfunctional.
Because high K+ induces vasoconstriction via effects on SMCs and nerve terminals, the neuronal contribution to this response can be evaluated by measuring the relaxation induced by blocking postjunctional α-adrenoreceptors by phentolamine. The phentolamine-induced relaxation of the contractile response to K+ was smaller in VEGF∂/∂ arteries (Table 2), suggesting that less noradrenaline reached and activated postjunctional α-adrenoreceptors. Overall, besides general SMC hyporesponsiveness, control of vasomotor tone by sympathetic nerves was defective in VEGF∂/∂ arteries. Vasorelaxation by afferent sensory-motor nerves18 was also impaired in VEGF∂/∂ arteries (Figure 3E), indicating that dysfunction of neuroeffector mechanisms involves both sympathetic and sensory-motor nerves.
Cellular Mechanisms of Neuroeffector Dysfunction in VEGF∂/∂ Mice
We then sought to explore the underlying mechanisms of the perivascular nerve dysfunction. Whole-mount staining for tyrosine hydroxylase, noradrenaline, or calcitonin-gene related peptide revealed that the neuroeffector dysfunction was not attributable to a reduced density of the periarterial nervous plexus (Figure IV of the online-only Data Supplement). Additionally, the noradrenaline content of VEGF∂/∂ resistance arteries was not different from that of controls (Table III of the online-only Data Supplement). We further analyzed by myography the effect of cocaine on the arterial response to exogenous noradrenaline to assess how efficiently noradrenaline reuptake by nerve terminals in situ occurred. Indeed, if noradrenaline uptake is important, blockade of this process by cocaine should leave more neurotransmitter available to activate postjunctional receptors at the NEJ. Notably, cocaine enhanced the sensitivity of VEGF∂/∂ arteries to noradrenaline, but much less than that of WT arteries (Table 2). Hyporesponsiveness to cocaine in this experimental condition might reflect impaired noradrenaline uptake per se but also could be due to a wider junctional cleft between nerve varicosities and SMCs because local neurotransmitter concentrations can be influenced more readily by uptake processes in narrow junctional clefts. To discriminate between these possibilities, we measured the neuronal uptake of noradrenaline directly. Notably, no differences in noradrenaline uptake between VEGF∂/∂ and WT arteries were found (note II in the online-only Data Supplement).
We therefore explored whether the junctional cleft between perivascular nerve varicosities and target vascular SMCs might be wider. To identify structural abnormalities of VEGF∂/∂ NEJs, transmission electron microscopy was used (Table 3 and Figure 3F and 3G). Morphometric quantification revealed that the distance between varicosities and the nearest SMC was larger in VEGF∂/∂ arteries (Table 3). Furthermore, varicosities were smaller and contained fewer mitochondria, and axon bundles contained fewer varicosities in VEGF∂/∂ arteries, whereas the density of transmitter storage vesicles was not altered (Table 3). Thus, neuroeffector dysfunction in VEGF∂/∂ arteries was accompanied by structural changes in NEJs in 3-month-old mice.
To assess whether these structural changes occurred during development or only after formation of the adult periarterial nervous network, we repeated the ultrastructural analysis on arteries of 3-week-old mice in which the periarterial nerve plexus had just matured.20 Even at this stage, the average distance between nerve varicosities and SMCs was increased in VEGF∂/∂ arteries, whereas the varicosity area, number of mitochondria per varicosity, and number of varicosities per nerve bundle were reduced in VEGF∂/∂ arteries (Table 3). Thus, low VEGF levels in VEGF∂/∂ mice perturb normal development of arterial NEJs.
Impaired Vascular Regulation of Resistance Arteries in VEGF∂/∂ Mice In Vivo
To assess whether the aforementioned ex vivo changes in neuroeffector function also impaired autonomic nervous control of peripheral resistance arteries in vivo, we explored whether thermoregulation, known to rely on autonomic nervous control of resistance arteries in the skin,3 was impaired in VEGF∂/∂ mice. In baseline conditions, the core body temperature was 38.2±0.1°C in adult WT mice but only 37.2±0.2°C in VEGF∂/∂ mice (n=8; P<0.005). When exposed to cold stress (8°C), VEGF∂/∂ mice were unable to maintain their body temperature and progressively cooled down to 34.2±1.1°C after 4.5 hours (n=8; P=0.002; Figure 4A). This thermoregulatory defect was not attributable to genotypic differences in blood pressure or heart rate (Table IV of the online-only Data Supplement), thyroid function (note III of the online-only Data Supplement), cutaneous vascular density (percent of surface area covered by vasculature: 29.3±1.4 in WT mice versus 29.7±1.0 in VEGF∂/∂ mice; n=3; P=0.83), skeletal muscle dysfunction, or brown fat tissue mass (not shown).
To further evaluate whether the ex vivo peripheral arterial defects were also relevant in vivo, we analyzed whether VEGF∂/∂ mice were capable of increasing cerebral blood flow (CBF) in conditions of general oxygen shortage. Indeed, when oxygen delivery to the brain becomes limiting, neurogenic constriction of resistance arteries in visceral organs ensures that brain perfusion is increased at the expense of peripheral organs.4 We therefore exposed anesthetized ventilated mice to a brief pulse of 8% oxygen (200 seconds) and measured changes in CBF using continuous arterial spin labeling. As expected, after the hypoxic challenge, CBF was increased in WT mice by 33% (Figure 4B). In contrast, the CBF in VEGF∂/∂ mice was increased by only 9% (n=4; P<0.01; Figure 4B). Thus, the ex vivo documented defects in vascular regulation of peripheral resistance arteries have in vivo physiological consequences.
Inhibition of VEGF Impairs Vascular Regulation in WT Mice
We then assessed whether the functional and structural defects of arterial NEJs could also be induced on short-term inhibition of VEGF in adult WT mice because this would provide insight into whether VEGF might be involved in the dynamic regulation of these junctions. We therefore hydroporated WT mice with a plasmid encoding sFlk1, which traps VEGF, or an empty plasmid as control. This strategy yields high plasma levels of this VEGF inhibitor for several weeks.17 Five days after sFlk1 gene transfer, plasma sFlk1 levels ranged from 500 to 4000 ng/mL, sufficient to inhibit VEGF-driven tumor growth.17 Notably, ELISA measurements, using mesenteric and saphenous arteries after saline perfusion to remove intravascular sFlk1, revealed that sFlk1 diffused into the arterial wall (note IV of the online-only Data Supplement).
To evaluate the functional consequences of short-term VEGF blockade, we analyzed whether sFlk1 gene transfer in WT mice (sFlk1 mice) induced the same phenotypic defects as observed in VEGF∂/∂ mice. Already 10 days after sFlk1 gene transfer, sFlk1 mice could not maintain their body temperature on exposure to cold stress (4°C during 4.5 hours; n=9 to 12; P<0.001; Figure 4C). In addition, the response of isolated saphenous arteries to EFS was impaired in sFlk1 mice, indicating that NEJs were dysfunctional (maximal contractile response expressed as percent of maximal response to noradrenaline: 43.0±9.8% in sFlk1 mice versus 79.8±15.6% in controls; n=7; P=0.0008; Figure 4D). However, different from VEGF∂/∂ mice, contractile SMC responses to high potassium, noradrenaline, or U46619 (Figure 4E through 4G) and expression of SMC markers (Table V of the online-only Data Supplement) were not altered in resistance arteries of sFlk1 mice. Control of vascular regulation by endothelial cells was also normal in sFlk1 mice (Figure V of the online-only Data Supplement). Ultrastructural analysis of mesenteric resistance arteries revealed an increase in the distance between neuronal varicosities and SMCs by ≈75% in sFlk1 arteries, as well as a reduction in the number of varicosities per axon bundle by ≈30% (Table VI of the online-only Data Supplement). The area of the varicosities and the number of mitochondria per varicosity were not significantly affected by sFlk1 (Table VI of the online-only Data Supplement). Together, these results show that even short-term treatment of adult mice with a VEGF inhibitor was sufficient to induce functional and structural changes in arterial NEJs without affecting SMC function.
The present study provides genetic evidence that subnormal VEGF levels in VEGF∂/∂ mice impair vascular regulation of resistance arteries, thereby hindering thermoregulation and a compensatory increase in cerebral perfusion in response to hypoxia. Pharmacological, expression, and morphological studies further revealed that this is due to dysfunction of arterial NEJs and to dedifferentiation of contractile SMCs. Moreover, short-term inhibition of VEGF in healthy mice caused neuroeffector dysfunction without SMC defects.
Neuroeffector dysfunction was not due to overt degeneration of periarterial nerves but rather to more subtle ultrastructural defects. Indeed, the junctional cleft between nerve varicosities and target vascular SMCs was widened in VEGF∂/∂ arteries, thereby lowering the fraction of released neurotransmitter, reaching and activating postjunctional receptors on target SMCs, and thus impairing neuroeffector function.21 In addition, perivascular nerves had smaller and fewer varicosities, and these varicosities contained fewer mitochondria. These defects may further contribute to the impaired neuroeffector function.
Vascular SMC dysfunction in VEGF∂/∂ mice was associated with reduced levels of late vascular SMC differentiation markers (smoothelin-B, SM-MHC), which are typically expressed by contractile arteries.19 So far, VEGF has been shown to stimulate SMC migration and to regulate the association of SMCs around endothelial cells22,23 but has not been reported to regulate vascular SMC differentiation. Our findings suggest that VEGF is required to maintain SMCs in a fully differentiated contractile state. Whether this activity of VEGF relies on a direct effect on SMCs or on an indirect effect through production of “arterializing” signals from the nerve terminal remains to be determined.
Several mechanistic models could explain the impaired vascular regulation of VEGF∂/∂ arteries (Figure 5). One model, which we favor at present, postulates that reduced release of VEGF by SMCs causes functional and structural defects of the NEJ through direct effects on the nerve terminal, which would secondarily cause SMC dedifferentiation (“prejunctional” mechanism). Arguments in favor of this model are that VEGF is expressed by vascular SMCs, that VEGF levels are reduced in VEGF∂/∂ arteries, and that VEGF receptors are expressed by periarterial nerves. A primary neuronal defect is also suggested by findings that neuroeffector defects are already present by 3 weeks of age, whereas SMC differentiation is still normal. Furthermore, sFlk1 gene transfer in adult WT mice caused structural and functional neuroeffector changes without inducing SMC alterations. Previous studies documented that cultured sympathetic neurons express VEGFRs and respond to VEGF by enhanced axon outgrowth and survival,12 but so far, VEGF has not been reported to induce structural changes in prejunctional nerve terminals. Because autonomic nerve endings are known to release SMC-trophic factors,24 a dysfunction of these nerve terminals may secondarily lead to SMC dedifferentiation. Indeed, activation of α1-adrenoreceptors by noradrenaline stimulates SMC proliferation,25 and sympathetic innervation increases vascular SMC contractile protein expression,24 whereas sympathectomy results in a switch from a contractile to synthetic SMC phenotype.26
An alternative nonexclusive model postulates that reduced release of VEGF from nerve endings and/or vascular SMCs could lead to SMC dedifferentiation, which would then secondarily cause defects in neuroeffector structure and function (“postjunctional” mechanism). Indeed, vascular SMCs are known to promote survival of cultured postganglionic sympathetic neurons.27 This model seems less likely because SMC differentiation is normal in young VEGF∂/∂ mice and adult sFlk1 mice in which neuroeffector defects are already present. Furthermore, quantitative polymerase chain reaction analysis revealed that the expression of nerve growth factor, artemin, and neurotrophin-3 (known to have plasticity effects on perivascular autonomic nerves28) was not reduced in VEGF∂/∂ arteries. Thus, we have little evidence to support a model whereby prejunctional nerve terminals, through paracrine release of VEGF, would ensure their own performance by inducing retrograde release of neurotrophic cues by target SMCs. We are therefore inclined to propose that VEGF may, at least in part, regulate the remodeling of the NEJ through direct neurotrophic effects on the prejunctional nerve terminals.
Short-term administration of the VEGF inhibitor sFlk1 to healthy adult mice was sufficient to cause vascular regulation defects and perivascular nerve dysfunction, indicating a role for VEGF in the maintenance of NEJs. An intriguing question is whether the observed functional and structural remodeling of the NEJs reflects a role for VEGF in regulating the “plasticity” of the perivascular autonomic nervous system and whether a similar role for VEGF in regulating “synaptic plasticity” exists in the central nervous system. Reports that VEGF stimulates long-term potentiation in the hippocampus would be consistent with such a model.29 These findings also raise the intriguing but unanswered question of whether impaired vascular regulation might explain, at least in part, the hand-foot syndrome in the treatment of cancer patients with VEGF (receptor) inhibitors.6 Indeed, impaired vascular regulation has been proposed to contribute to the hand-foot syndrome because dilated blood vessels are commonly found in affected tissues,8 and decreasing blood flow by cooling hands and feet lowers the incidence of the hand-foot syndrome.30 Overall, our genetic and pharmacological studies uncover a previously unrecognized role of VEGF in the proper functioning and dynamic remodeling of perivascular autonomic NEJs.
We thank G. van Eys for providing the smoothelin antibody. We also thank H. Bellens, A. Bouché, N. Dai, M. Deprez, M. De Mol, K. Feyen, B. Hermans, S. Jansen, L. Kieckens, A. Manderveld, M. Nijs, W. Martens, L. Notebaert, V. Uytterhoeven, M. Vanbrabant, A. Van Den Broeck, B. Vanwetswinkel, E. Weltens, and S. Wyns for their contributions.
Sources of Funding
Drs Storkebaum and Ruiz de Almodovar are postdoctoral fellows of Fonds voor Wetenschappelijk Onderzoek (FWO)–Flanders. This work was supported by Muscular Dystrophy Association grant 3751, FWO grant G.0113.02, long-term structural funding from Methusalem funding by the Flemish Government, European Union grant QLK6-CT-2000-5303, Geconcerteerde Onderzoeksacties grant GOA2006/11, a grant from the Transnational University of Limburg, Top Institute Pharma grant TIP2_108, a European Vascular Genomics Network grant, European Commission FP6-projects DiMI (LSHB-CT-2005-512146) and EMIL (LSHC-CT-2004-503569), and an Agentschap voor Innovatie door Wetenschap en Technologie SBO grant.
Romanovsky AA. Thermoregulation: some concepts have changed: functional architecture of the thermoregulatory system. Am J Physiol Regul Integr Comp Physiol. 2007; 292: R37–R46.
Kuwahira I, Gonzalez NC, Heisler N, Piiper J. Changes in regional blood flow distribution and oxygen supply during hypoxia in conscious rats. J Appl Physiol. 1993; 74: 211–214.
Oosthuyse B, Moons L, Storkebaum E, Beck H, Nuyens D, Brusselmans K, Dorpe JV, Hellings P, Gorselink M, Heymans S, Theilmeier G, Dewerchin M, Laudenbach V, Vermylen P, Raat H, Acker T, Vleminckx V, Bosch LV, Cashman N, Fujisawa H, Drost MR, Sciot R, Bruyninckx F, Hicklin DJ, Ince C, Gressens P, Lupu F, Plate KH, Robberecht W, Herbert JM, Collen D, Carmeliet P. Deletion of the hypoxia-response element in the vascular endothelial growth factor promoter causes motor neuron degeneration. Nat Genet. 2001; 28: 131–138.
Sondell M, Lundborg G, Kanje M. Vascular endothelial growth factor has neurotrophic activity and stimulates axonal outgrowth, enhancing cell survival and Schwann cell proliferation in the peripheral nervous system. J Neurosci. 1999; 19: 5731–5740.
Marko SB, Damon DH. VEGF promotes vascular sympathetic innervation. Am J Physiol Heart Circ Physiol. 2008; 294: H2646–H2652.
Fischer C, Jonckx B, Mazzone M, Zacchigna S, Loges S, Pattarini L, Chorianopoulos E, Liesenborghs L, Koch M, De Mol M, Autiero M, Wyns S, Plaisance S, Moons L, van Rooijen N, Giacca M, Stassen JM, Dewerchin M, Collen D, Carmeliet P. Anti-PlGF inhibits growth of VEGF(R)-inhibitor-resistant tumors without affecting healthy vessels. Cell. 2007; 131: 463–475.
De Mey JG, Megens R, Fazzi GE. Functional antagonism between endogenous neuropeptide Y and calcitonin gene-related peptide in mesenteric resistance arteries. J Pharmacol Exp Ther. 2008; 324: 930–937.
Owens GK. Regulation of differentiation of vascular smooth muscle cells. Physiol Rev. 1995; 75: 487–517.
Wang H, Keiser JA. Vascular endothelial growth factor upregulates the expression of matrix metalloproteinases in vascular smooth muscle cells: role of flt-1. Circ Res. 1998; 83: 832–840.
Damon DH. Sympathetic innervation promotes vascular smooth muscle differentiation. Am J Physiol Heart Circ Physiol. 2005; 288: H2785–H2791.
van Kleef EM, Smits JF, De Mey JG, Cleutjens JP, Lombardi DM, Schwartz SM, Daemen MJ. Alpha 1-adrenoreceptor blockade reduces the angiotensin II-induced vascular smooth muscle cell DNA synthesis in the rat thoracic aorta and carotid artery. Circ Res. 1992; 70: 1122–1127.
Damon DH. NGF-independent survival of postganglionic sympathetic neurons in neuronal-vascular smooth muscle cocultures. Am J Physiol Heart Circ Physiol. 2001; 280: H1722–H728.
Zimmerman GC, Keeling JH, Lowry M, Medina J, Von Hoff DD, Burris HA. Prevention of docetaxel-induced erythrodysesthesia with local hypothermia. J Natl Cancer Inst. 1994; 86: 557–558.
Vascular endothelial growth factor (VEGF) is a key angiogenic factor, and blocking of VEGF signaling in cancer patients inhibits angiogenesis, thereby preventing tumor progression. One of the side effects of anti-VEGF therapy is the hand-foot syndrome, which is characterized by dilated blood vessels in hands and feet, leading to dysesthesia, erythema, and tingling of extremities, which may progress to burning pain with dryness, cracking, and ulceration of the skin. Here, we show that knock-in mice with reduced VEGF levels display abnormal regulation of resistance arteries, resulting in defects in thermoregulation and redistribution of blood flow in response to hypoxia. The abnormal vascular regulation was attributed to reduced vascular smooth muscle cell contractility associated with reduced expression of contractile proteins, as well as to dysfunction of neuroeffector junctions as a result of their abnormal structural development. Furthermore, short-term treatment of wild-type mice with a VEGF trap also induced thermoregulation defects and neuroeffector dysfunction with structural remodeling of neuroeffector junctions, indicating that VEGF is necessary for maintenance of the structural and functional integrity of adult neuroeffector junctions. These findings may provide mechanistic insights into the cause of the hand-foot syndrome.
↵*Drs Storkebaum and Ruiz de Almodovar contributed equally to this work.
The online-only Data Supplement is available with this article at http://circ.ahajournals.org/cgi/content/full/CIRCULATIONAHA.109.929364/DC1.