Author + information
- Received August 20, 2007
- Revision received December 28, 2007
- Accepted February 7, 2008
- Published online July 29, 2008.
- Clarence M. Findley, PhD⁎,†,‡,
- Robert G. Mitchell, MD⁎,
- Brian D. Duscha, MS⁎,
- Brian H. Annex, MD⁎,§ and
- Christopher D. Kontos, MD, FACC⁎,‡,⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. Christopher D. Kontos, Box 3629 DUMC, Duke University Medical Center, Durham, North Carolina 27710.
Objectives Our purpose was to determine whether factors that regulate angiogenesis are altered in peripheral arterial disease (PAD) and whether these factors are associated with the severity of PAD.
Background Alterations in angiogenic growth factors occur in cardiovascular disease (CVD), but whether these factors are altered in PAD or correlate with disease severity is unknown.
Methods Plasma was collected from patients with PAD (n = 46) and healthy control subjects (n = 23). Peripheral arterial disease patients included those with intermittent claudication (IC) (n = 23) and critical limb ischemia (CLI) (n = 23). Plasma angiopoietin-2 (Ang2), soluble Tie2 (sTie2), vascular endothelial growth factor (VEGF), soluble VEGF receptor 1 (sVEGFR-1), and placenta growth factor (PlGF) were measured by enzyme-linked immunoadsorbent assay. In vitro, endothelial cells (ECs) were treated with recombinant VEGF to investigate effects on sTie2 production.
Results Plasma concentrations of sTie2 (p < 0.01), Ang2 (p < 0.001), and VEGF (p < 0.01), but not PlGF or sVEGFR-1, were significantly greater in PAD patients compared with control subjects. Plasma Ang2 was significantly increased in both IC and CLI compared with control subjects (p < 0.0001), but there was no difference between IC and CLI. Plasma VEGF and sTie2 were similar in control subjects and IC but were significantly increased in CLI (p < 0.001 vs. control or IC). Increased sTie2 and VEGF were independent of CVD risk factors or the ankle-brachial index, and VEGF treatment of ECs in vitro significantly increased sTie2 shedding.
Conclusions Levels of VEGF and sTie2 are significantly increased in CLI, and sTie2 production is induced by VEGF. These proteins may provide novel biomarkers for CLI, and sTie2 may be both a marker and a cause of CLI.
- peripheral arterial disease
- soluble Tie2
- vascular endothelial growth factor
- critical limb ischemia
Peripheral arterial disease (PAD) is characterized by atherosclerosis of noncardiac vascular beds and most commonly affects the lower extremities. Although widely under-recognized, PAD is a major health problem, affecting 8 to 12 million individuals in the U.S. (1). The 2 major clinical manifestations of PAD of the lower extremities are intermittent claudication (IC) and critical limb ischemia (CLI), which have markedly different clinical outcomes (2). Intermittent claudication is characterized by reproducible pain on exertion that is relieved with rest. Critical limb ischemia is the most severe form of PAD and is characterized by the inability of arterial blood flow to meet the metabolic demands of resting muscle or tissue, resulting in rest pain and/or tissue necrosis, which frequently necessitates amputation (3,4). Whereas annual mortality in patients with IC is only 1% to 2%, the annual risk of death in patients with CLI is up to 20% (3,4). Therefore, early identification would be expected to provide improved treatment options for patients with PAD.
Currently, the ankle-brachial index (ABI) is the standard for diagnosis of PAD, however the correlation between ABI and severity of PAD is poor (5). Aside from the ABI, there are no other reliable diagnostic tests for PAD. Furthermore, the diagnosis of IC versus CLI is purely clinical. Patients with IC and CLI can present with virtually identical clinical risk factors (e.g., tobacco use, diabetes mellitus), peripheral hemodynamic parameters, and degree of atherosclerotic burden (6). Therefore, minimally invasive tests or biomarkers to diagnose PAD and potentially distinguish between IC and CLI are needed.
In PAD, atherosclerotic arterial occlusive disease results in tissue ischemia and varying degrees of collateral blood vessel growth (i.e., angiogenesis). The extent of collateral blood vessel formation has the potential to affect the clinical manifestations of PAD; thus insufficient angiogenesis may be responsible for the different clinical presentations of patients with IC and CLI. A number of angiogenic growth factors and endogenous inhibitors of angiogenesis have the potential to modulate vascular growth in PAD. Among these, vascular endothelial growth factor (VEGF)-A is perhaps the most important proangiogenic factor, because it is required for both physiological and pathological angiogenesis (7) and even minor changes in its expression can lead to dramatic alterations in vascular growth and remodeling (8). The related placenta growth factor (PlGF) binds a common receptor, VEGF receptor 1 (VEGFR-1), and although it is not required for normal embryonic vascular development, it plays a critical role in pathological angiogenesis (9). The angiopoietins are ligands for the endothelial receptor Tie2 (10), and angiopoietin-2 (Ang2) is up-regulated by VEGF (11) and in some cases has been shown to be required for VEGF-mediated angiogenesis (12). Interestingly, soluble forms of both VEGFR-1 and Tie2 are produced and secreted into the circulation, where they can bind their respective growth factors and inhibit angiogenesis (13). Whereas soluble VEGFR-1 (sVEGFR-1) is produced by alternative splicing of the VEGFR-1 gene (14), soluble Tie2 (sTie2) results from proteolytic cleavage of the full-length membrane-bound receptor (15), and both proteins have been implicated in the pathogenesis of cardiovascular disease (CVD) (16,17).
For the present report, we hypothesized that differences in expression of these angiogenic growth factors and/or their soluble receptors might provide markers of PAD and its different clinical manifestations. To investigate this possibility, plasma from control subjects and patients with IC and CLI was analyzed for changes in the proangiogenic growth factors Ang2, VEGF-A, and PlGF, along with their soluble receptors sVEGFR-1 and sTie2, which act as endogenous inhibitors of angiogenesis. We found that patients with PAD have significantly higher plasma concentrations of Ang2, VEGF, and sTie2 compared with control subjects, but only VEGF and sTie2 concentrations distinguished patients with CLI from those with IC.
Subjects with PAD were recruited consecutively from the vascular medicine clinics at Duke University Medical Center, and control subjects were recruited from the vascular screening clinics if they showed no evidence of PAD, as demonstrated by an ABI of >1.0 or by other diagnostic testing. All patient studies were approved by the Duke University Institutional Review Board. All patients were over 18 years of age and able to give informed consent for a single 10-ml blood draw. The study population comprised 23 control subjects without PAD and 46 patients with PAD. Peripheral arterial disease subjects included 23 patients with IC and 23 patients with CLI, based on the Rutherford criteria (18). Intermittent claudication patients had an ABI of 0.5 to 0.9 and exercise-limiting claudication, and the CLI subjects were identified by an ABI of <0.5 and the presence of a nonhealing lower extremity ulcer, rest pain, or gangrene. Control subjects were age and gender matched and were without clinical evidence of vascular, metabolic, or neoplastic disease, based on patient history and physical examination and routine laboratory studies. Subjects were excluded if they had evidence of acute coronary syndromes and/or acute decompensated heart failure in the preceding 6 weeks or Child-Pugh class C liver disease. Sample size was determined based on 80% power to detect a significant difference in at least 1 measured factor between control and PAD subjects without distinction between IC and CLI subgroups. The subjects' demographic characteristics are shown in Table 1.
Sample acquisition and analysis of angiogenesis-modulating factors
Blood was collected by venipuncture into EDTA-containing tubes. Whole blood samples were placed on ice and immediately centrifuged at 2,000 × g for 5 min at 4°C for plasma separation. Plasma was aliquoted and stored at −80°C. The concentrations of sTie2, VEGF-A165, Ang2, PlGF, and sVEGFR-1 were quantified using commercially available enzyme-linked immunoadsorbent assay (ELISA) kits (R&D Systems, Minneapolis, Minnesota) according to the manufacturer's instructions. Assay sensitivities, quantified as the mean minimal detectable concentration for each assay, were: sTie2 14 pg/ml; VEGF-A 5.0 to 9.0 pg/ml; Ang2 8.3 pg/ml; PlGF 7 pg/ml; and sVEGFR-1 3.5 pg/ml. Coefficients of variation (intra-assay precision) were <4% for all assays.
Cell culture studies
Human umbilical vein endothelial cells (HUVECs) were harvested from fresh umbilical cords, phenotyped as described previously (19), and used between passages 2 and 6. Cells were grown in endothelial growth medium (EGM)-MV (Cambrex, Walkersville, Maryland) until confluent. The cells were then changed to serum-free medium, and triplicate samples were treated with vehicle or with recombinant human VEGF165 (25 ng/ml, R&D Systems) for 24 h. The conditioned media were then collected and sTie2 concentration was quantified by ELISA, as described in the preceding text. Enzyme-linked immunoadsorbent assay data are presented as the mean ± SD from triplicate samples. Experiments were repeated on at least 3 separate occasions. Statistical comparisons were performed using Student t test, and significance was set at p < 0.05.
Statistical analysis was performed using SPSS 16.0 for Windows (SPSS Inc., Chicago, Illinois). Patient demographic data were compared across all 3 groups by analysis of variance (ANOVA) for continuous variables followed by chi-square analysis to compare IC and CLI patients. Categorical variables were compared between IC and CLI groups using a nonparametric chi-square test. Categorical variables that were significantly different between the IC and CLI patients were controlled for using analysis of covariance (ANCOVA) to compare levels of VEGF and sTie2 within the PAD cohort. Similarly, ANCOVA was used to control for the effects of differences in ABI on VEGF and sTie2 concentrations in IC versus CLI patients. For analysis of ELISA data, which were not normally distributed, the data were log-transformed and analyzed by 1-way ANOVA followed by the Fisher least significant difference test or the Student t test where appropriate. Graphical data are presented as the non–log-transformed values and are expressed as the mean ± SD. Statistical significance was set at p < 0.05.
Baseline patient characteristics
The baseline demographic and clinical characteristics of the study population are shown in Table 1. Comparing patients with IC and CLI, a significantly greater percentage with CLI were male, diabetic, hypertensive, and had a history of smoking and peripheral vascular intervention (Table 1). There were no significant differences in the use of cardiovascular medications (angiotensin-converting enzyme inhibitor/angiotensin receptor blocker or statins) between patients with IC and those with CLI (Table 1).
Analysis of angiogenic growth factors and inhibitors in PAD and control subjects
Plasma concentrations of the 5 different factors were analyzed in control subjects versus all PAD patients combined. Plasma concentrations of sVEGFR-1 and PlGF were not significantly different in PAD patients compared with control subjects (Figs. 1A and 1B). However, levels of Ang2, VEGF, and sTie2 were significantly higher in PAD patients (Figs. 1C to 1E).
Analysis of angiogenic growth factors and inhibitors within PAD groups (IC vs. CLI)
Visually, there was a stepwise increase in plasma Ang2 expression from control subjects to patients with IC to patients with CLI (Fig. 2A). Although Ang2 concentrations were significantly greater in both PAD groups compared with controls, Ang2 expression was not statistically different in patients with IC and those with CLI (Fig. 2A). In contrast, plasma VEGF expression was not significantly increased in IC patients compared with control subjects (Fig. 2B). However, plasma VEGF expression was significantly higher in patients with CLI compared with those with IC (Fig. 2B). Similarly, plasma sTie2 levels were not significantly different in patients with IC compared with controls, but they were significantly higher in CLI patients compared with either IC patients or control subjects (Fig. 2C). Mean sTie2 levels in the 3 groups were 20.6 ± 1.8 ng/ml, 22.8 ± 1.5 ng/ml, and 32.1 ± 1.4 ng/ml in control, IC, and CLI patients, respectively.
sTie2 and VEGF concentrations are independent of cardiovascular risk factors
Circulating concentrations of VEGF and sTie2 could be altered as a result of concomitant CVD and endothelial dysfunction. Analysis of the baseline demographic and clinical characteristics of patients with IC and CLI revealed that significantly more CLI patients were male, diabetic, hypertensive, and had a history of smoking (Table 1). Therefore, we performed ANCOVA for each of these categorical variables with VEGF and sTie2 concentrations and found that differences in both VEGF and sTie2 between the IC and CLI groups remained significant (p < 0.001) in the present study population.
sTie2 and VEGF concentrations are independent of the ABI
The ABI values were lower within the CLI patients than in those with IC (Table 1). However, when we controlled for differences in ABI, the increased concentrations of both sTie2 and VEGF in CLI versus IC patients remained significantly different (p < 0.001). These findings demonstrate that increased levels of sTie2 and VEGF correlate with disease state but not with the ABI, suggesting that these proteins are potential markers of CLI.
Interaction between VEGF and sTie2
The finding that sTie2 and VEGF concentrations were similarly elevated in subjects with CLI compared with controls or those with IC suggested a potential mechanistic link between VEGF and sTie2. It is known that VEGF activates proteases upon stimulation of endothelial cells (7) and that sTie2 is a proteolytic cleavage product of the full-length membrane-bound receptor (20). Moreover, VEGF has been shown to induce proteolysis of the related Tie1 receptor (21). Therefore, we stimulated HUVECs with recombinant human VEGF and analyzed cell-conditioned medium for the presence of cleaved sTie2 by ELISA. Vascular endothelial growth factor induced a significant increase in sTie2 in vitro (Fig. 3), suggesting that increased sTie2 levels in CLI are a result of increased VEGF expression.
Peripheral arterial disease is a major health care problem in the U.S., and the prevalence of PAD is increasing. Although the ABI is currently the gold standard for the diagnosis of PAD, there are a number of limitations to its use, and new diagnostic tests are needed. Because PAD is characterized by decreased tissue perfusion and tissue hypoxia, we hypothesized that concentrations of circulating angiogenic growth factors and/or inhibitors of angiogenesis would be altered in subjects with PAD and that these factors might provide important diagnostic and therapeutic targets in this disease. We found that plasma concentrations of Ang2, VEGF, and sTie2 were all significantly elevated in patients with PAD compared with control subjects. Furthermore, plasma concentrations of VEGF and sTie2, but not Ang2, were significantly increased in patients with CLI compared with those with IC. Increases in VEGF and sTie2 in CLI patients were independent of CVD risk factors and the ABI. Moreover, VEGF was found to induce shedding of sTie2 from endothelial cells in vitro, suggesting a potential mechanistic link between these 2 proteins. Taken together, these findings suggest that sTie2 and VEGF may provide novel biomarkers for PAD in general and CLI more specifically. These results have implications for understanding the pathophysiology and potentially improving the diagnosis and treatment of PAD.
Despite its increasing prevalence, PAD continues to be clinically under-recognized. Recent data indicate that a significant percentage of patients with PAD are asymptomatic or present with atypical symptoms (22). Moreover, patients with PAD frequently have concomitant coronary artery disease, and their risk of cardiovascular events is significantly higher than in patients without PAD. As a result, the early identification of PAD is of paramount importance, because it would lead to earlier and more aggressive cardiovascular risk factor modification. The identification of novel biomarkers of PAD would aid substantially in this effort. To this end, Wilson et al. (23) recently used an unbiased proteomic approach to demonstrate that β2-microglobulin (β2M) is up-regulated in the serum of patients with PAD. β2-microglobulin is a component of the class I major histocompatibility complex and is present on virtually all cell membranes, and although the mechanism of its increased release is not understood, its identification as a potential biomarker of PAD represents an important step forward in the diagnosis and management of this disease (24). Proteomic techniques such as the mass spectroscopy approach used by Wilson et al. (23) provide an attractive means of identifying potential protein biomarkers. However, an important limitation of these approaches is that they can detect only a small subset of the total serum proteome. Many serum proteins are masked by albumin or other abundant proteins, and a number of proteins of interest, including most cytokines and growth factors, are present in serum at concentrations that are below the level of detection with these techniques. Therefore, the hypothesis-driven targeted approach used in the present study provides an important complement in investigating potential biomarkers for PAD.
Peripheral arterial disease is characterized by tissue ischemia, and the resulting tissue hypoxia provides a potent angiogenic stimulus. Therefore, PAD should result in up-regulation of angiogenic factors in order to induce a compensatory angiogenic response, and VEGF, Ang2, and sVEGFR-1 are all known to be up-regulated by hypoxia in vitro (25–27). A number of studies have examined circulating levels of angiogenesis-modulating proteins in the setting of CVD or in patients with CVD risk factors, such as hypertension and diabetes mellitus. Similar to our results, plasma concentrations of VEGF, Ang2, and sTie2 have been found to be increased the setting of coronary artery disease, acute coronary syndromes (17,28), and hypertension (29). Increased VEGF and Ang2 have also been demonstrated in the setting of diabetes (30), although there appear to be no data on sTie2 levels in diabetics. In the context of PAD, only VEGF and sVEGFR-1 have been studied previously, and, consistent with our findings, plasma VEGF concentrations were found to be higher in patients with PAD compared with controls (31–33). Also in agreement with our results, circulating levels of sVEGFR-1 were not increased in patients with PAD but were either unchanged (31,33) or reduced (32) compared with controls. Importantly, after controlling for major demographic and cardiovascular risk factors in the present study, the differences in circulating sTie2 and VEGF levels in CLI patients remained significant.
Our study subdivided PAD patients by their diagnosis of IC or CLI, which have markedly different outcomes. Currently, the diagnosis of IC versus CLI is largely clinical and is based on a combination of signs, symptoms, and, to a certain degree, the ABI. Our finding that VEGF and sTie2 were significantly increased in patients with CLI compared to IC indicates that plasma concentrations of these proteins may distinguish between these two conditions. To our knowledge, only one study has attempted to correlate expression of angiogenic proteins with severity of PAD. Makin et al. (31) examined VEGF levels in PAD patients as a function of the ABI and found that VEGF concentrations did not differ in patients with ABIs above or below 0.52. Although that study did not classify patients clinically as having CLI, no correlation was found between plasma VEGF concentration and the presence of rest pain (31). A critical difference in our study is that patients were given a clinical diagnosis of CLI that was not based on the ABI alone, since it correlates poorly with disease state. Importantly, our data demonstrate a significant association between levels of sTie2 or VEGF and the clinical diagnosis of CLI that is independent of the ABI. These findings suggest that sTie2 and/or VEGF may provide novel biomarkers for this manifestation of PAD.
Whereas VEGF is likely up-regulated as a compensatory response to ischemia in PAD, sTie2 appears more likely to play a role in the pathogenesis of the disease. Although IC and CLI have traditionally been thought of as early versus late or minor versus severe variants of the same disease, recent evidence suggests that these manifestations of PAD may be caused by distinct mechanisms. In this regard, the up-regulation of an inhibitor of angiogenesis such as sTie2 could worsen the clinical presentation despite having a similar extent of vascular occlusive disease. If so, sTie2 could serve as a target for therapeutic intervention in PAD. Furthermore, the increase in circulating levels of both VEGF and sTie2 in CLI patients suggests a mechanistic link between these 2 proteins. Accordingly, the present data demonstrated that VEGF induced proteolytic cleavage and shedding of sTie2 in vitro, and recent concomitant work from our laboratory has demonstrated that this process is mediated through the phosphoinositide 3-kinase/Akt pathway (15). Therefore, it is possible that ischemia-induced increases in VEGF expression result in dysregulated cleavage of sTie2, thereby limiting the angiogenic response to tissue hypoxia in those patients with the most severe ischemia, eventually resulting in CLI.
The regulation of Tie2 cleavage and subsequent shedding of sTie2 is poorly understood. Earlier studies have demonstrated sTie2 in the serum of healthy individuals (20) as well as in patients with a variety of CVDs, including congestive heart failure, hypertension, and acute coronary syndromes (17,29,34). Moreover, in renal cell carcinoma patients, sTie2 concentrations correlated with severity of disease and decreased survival (35). The present study is the first to examine sTie2 levels in PAD, and the correlation of sTie2 with CLI provides important insights into this manifestation of PAD.
Although a clear limitation of the current study is the modest sample size, a unique feature of the study is the analysis of subclassifications of patients with PAD (i.e., IC and CLI). Moreover, investigation of changes in factors involved in angiogenesis provides novel insights into the pathophysiology of this disease and its different manifestations. Although the present results suggest that plasma concentrations of Ang2 are not significantly different between patients with IC and those with CLI, the study was underpowered to rule out a type II error (i.e., a false negative) in this analysis. Therefore, assessment of larger populations will be necessary to confirm whether Ang2 might also serve as a marker of CLI as well as to validate the present observations on sTie2 and VEGF in this population. Furthermore, investigation of temporal changes in these factors in PAD populations may be clinically relevant, and additional studies will be needed to establish quantitative values of sTie2 and/or VEGF that define CLI with relatively high sensitivity and specificity. The targeted hypothesis-driven approach used here to investigate changes in angiogenesis modulatory proteins is limited in its analysis of only a small subset of potential biomarkers of PAD. Therefore, additional unbiased proteomic approaches, as described recently for the identification of β2M (23), might lead to the identification of additional PAD biomarkers.
The present study demonstrated that plasma levels of sTie2 and VEGF are significantly increased in patients with PAD and that increases in these proteins distinguish patients with CLI from those with IC. Because the signs and symptoms of CLI can overlap with those of other disease states, including neuropathies and venous ulcers, these results could have important implications for the diagnosis of CLI. These results also demonstrate a potential mechanistic link between VEGF and sTie2 in PAD, suggesting that sTie2 may serve not only as a marker of CLI but also as a viable target for therapeutic intervention in this disease.
Supported in part by National Institutes of Health grants R01HL70165 and R21DK069673 (to Dr. Kontos), R01 HL075752 (to Dr. Annex), and R36AG027584 (to Dr. Findley); by a Grant-in-Aid (0655493U) from the Mid-Atlantic Affiliate of the American Heart Association (to Dr. Kontos); and by a grant from Medtronic, Inc. (to Dr. Mitchell). Dr. Findley was supported in part by a Fellowship Award from the UNCF-Merck Foundation. Drs. Findley and Mitchell contributed equally to this work.
Presented in part at the 17th Annual Scientific Sessions of the Society for Vascular Medicine and Biology, Philadelphia, Pennsylvania, June 2–3, 2006.
- Abbreviations and Acronyms
- ankle-brachial index
- critical limb ischemia
- cardiovascular disease
- human umbilical vein endothelial cell
- intermittent claudication
- peripheral arterial disease
- placenta growth factor
- soluble Tie2
- soluble vascular endothelial growth factor receptor 1
- vascular endothelial growth factor
- Received August 20, 2007.
- Revision received December 28, 2007.
- Accepted February 7, 2008.
- American College of Cardiology Foundation
- Criqui M.H.
- Dolan N.C.,
- Liu K.,
- Criqui M.H.,
- et al.
- Weitz J.I.,
- Byrne J.,
- Clagett G.P.,
- et al.
- Oh H.,
- Takagi H.,
- Suzuma K.,
- Otani A.,
- Matsumura M.,
- Honda Y.
- Siemeister G.,
- Schirner M.,
- Weindel K.,
- et al.
- Barleon B.,
- Siemeister G.,
- Martinybaron G.,
- Weindel K.,
- Herzog C.,
- Marme D.
- Findley C.M.,
- Cudmore M.J.,
- Ahmed A.,
- Kontos C.D.
- Hazarika S.,
- Dokun A.O.,
- Li Y.,
- Popel A.S.,
- Kontos C.D.,
- Annex B.H.
- Lee K.W.,
- Lip G.Y.H.,
- Blann A.D.
- Yabkowitz R.,
- Meyer S.,
- Black T.,
- Elliott G.,
- Merewether L.A.,
- Yamane H.K.
- Wilson A.M.,
- Kimura E.,
- Harada R.K.,
- et al.
- Annex B.H.
- Pichiule P.,
- Chavez J.C.,
- LaManna J.C.
- Gerber H.P.,
- Condorelli F.,
- Park J.,
- Ferrara N.
- Nadar S.K.,
- Blann A.,
- Beevers D.G.,
- Lip G.Y.H.
- Lim H.S.,
- Blann A.D.,
- Chong A.Y.,
- Freestone B.,
- Lip G.Y.
- Chong A.Y.,
- Caine G.J.,
- Freestone B.,
- Blann A.D.,
- Lip G.Y.
- Harris A.L.,
- Reusch P.,
- Barleon B.,
- Hang C.,
- Dobbs N.,
- Marme D.