Noninvasive Imaging of Angiogenesis With a 99mTc-Labeled Peptide Targeted at αvβ3 Integrin After Murine Hindlimb Ischemia
Background— Noninvasive imaging strategies play a critical role in assessment of the efficacy of angiogenesis therapies. The αvβ3 integrin is activated in angiogenic vessels and represents a potential target for noninvasive imaging of angiogenesis.
Methods and Results— We evaluated a 99mTc-labeled peptide (NC100692) targeted at αvβ3 integrin for imaging in an established murine model of angiogenesis induced by hindlimb ischemia. Control mice (n=9) or mice with surgical right femoral artery occlusion (n=29) were injected with NC100692 (1.5±0.2 mCi IV) at different times after femoral occlusion (1, 3, 7, and 14 days) for in vivo pinhole planar gamma camera imaging. Tissue from hindlimb proximal and distal to occlusion was excised for gamma well counting and for immunostaining. On in vivo pinhole images, increased focal NC100692 activity was seen distal to the occlusion at days 3 and 7. This increase in relative NC100692 activity was confirmed by gamma well counting. Lectin staining confirmed increased angiogenesis in the ischemic hindlimb at these time points. A fluorescent analogue of NC100692 was used to confirm specificity and localization of the targeted tracer in cultured endothelial cells. In addition, endothelial cell specificity was confirmed on tissue sections with the use of dual immunofluorescent staining of endothelium and the fluorescent analogue targeted at the αvβ3 integrin.
Conclusions— A 99mTc-labeled peptide (NC100692) targeted at αvβ3 integrin selectively localized to endothelial cells in regions of increased angiogenesis and could be used for noninvasive serial “hot spot” imaging of angiogenesis. This targeted radiotracer imaging approach is a major advance in tracking therapeutic myocardial angiogenesis and has an important clinical potential.
Received June 17, 2004; revision received February 4, 2005; accepted February 16, 2005.
Angiogenesis represents the formation of new capillaries by cellular outgrowth from existing microvessels and plays an important role in response to ischemia associated with peripheral arterial occlusive disease and myocardial infarction.1,2 A noninvasive approach for evaluation of angiogenesis would be valuable in risk stratification of patients with peripheral or myocardial ischemia and for the assessment of response to therapies directed at stimulation of angiogenesis or arteriogenesis. Most imaging approaches involve indirect evaluation of the physiological consequences of ischemia or therapeutic interventions.3 However, a more sensitive noninvasive approach would involve targeted imaging of the mediators of these processes. Angiogenesis is mediated by a number of growth factors and signaling proteins and involves the interaction of endothelial cells and the extracellular matrix.4 The αvβ3 integrin is one of the key cell surface receptors and adhesion molecules in angiogenesis and is upregulated in the setting of angiogenesis.5 We recently demonstrated that angiogenesis in the heart can be imaged noninvasively with an 111In-labeled peptidomimetic targeted at the αvβ3 integrin.6 We hypothesized that the angiogenic process associated with peripheral vascular disease can also be directly and noninvasively tracked with the use of a radiolabeled peptide targeted at the αvβ3 integrin.
See p 3188
The present study evaluated a 99mTc-labeled peptide (NC100692, Amersham Health) targeted at the αvβ3 integrin for noninvasive detection of angiogenesis associated with peripheral limb ischemia. NC100692 is a chelate-peptide conjugate containing an RGD (Arg-Gly-Asp) motif in a configuration that allows high affinity (Ki ≈1 nmol/L) and specific binding to the αvβ3 integrin.7 This agent has demonstrated acceptable radiolabeling, pharmacokinetics, and in vivo efficacy for imaging of tumor angiogenesis. We evaluated the serial changes in the retention of NC100692 using high-resolution gamma camera imaging in an established murine model of hindlimb ischemia–induced angiogenesis. This noninvasive approach for targeted αvβ3 integrin radiotracer imaging was validated by gamma well counting of tissue, and the endothelial cell specificity was confirmed by immunostaining.
Murine microvascular endothelial cells were stained with a fluorescein-labeled homologue of NC100692 and a control non-RGD peptide (Arg-Gly-Phe [GDF]; Amersham Health) with the use of methods previously described.8
Animal Preparation and Model
All experiments were performed with approval of the institutional Animal Care and Use Committee, according to the principles of the American Physiological Society on research animal use.
Male C57BL/6 mice (Charles River Laboratories, Wilmington, Mass) were anesthetized with ketamine (100 mg/kg IP) and xylazine (5 mg/kg IM) for imaging and surgical intervention. A subset of mice (n=29) underwent surgical occlusion to induce unilateral hindlimb ischemia according to the procedure described elsewhere.9,10 Briefly, the right femoral artery was exposed, and 2 ligatures were placed distal to the profundus branch, creating distal hindlimb ischemia. The left hindlimb underwent a sham operation involving the overlying skin and subcutaneous tissue. Mice were allowed to recover for 1 to 14 days after surgery. Mice with femoral occlusion were divided into 4 groups on the basis of the time of euthanasia after femoral artery occlusion (days 1, 3, 7, and 14). Another group of mice (n=9) did not undergo femoral ligation and served as a control.
Pinhole Planar Imaging
All animals were injected intravenously with 1.5±0.2 mCi of 99mTc-NC100692 targeted at the αvβ3 integrin. Mice with femoral occlusion were injected on days 1, 3, 7, and 14 of hindlimb ischemia. Animals were placed on a polyacrylic board in the supine position with legs secured at defined markers in an extended position for each image acquisition. This approach ensured a uniform orientation during each imaging session. Mice underwent early dynamic pinhole planar imaging, followed by a static 15-minute image at 75 minutes after injection of the radiotracer. Planar imaging was performed with a large field of view gamma camera (VariCam, GE Healthcare) and a pinhole collimator with a 1-mm aperture with a 15% energy window centered at 140 keV. The mice were euthanized immediately after the static imaging was completed. Tissue samples were taken for gamma well counting and immunostaining.
The acquired images were analyzed with the use of rectangular regions of interest placed over the ischemic hindlimb above and below the site of femoral artery occlusion. The standardized site of occlusion was identified on a set of reference images acquired with small 99mTc-labeled point sources placed at known distances from the ligature. Identical regions of interest were flipped horizontally and placed on the contralateral control limb. The maximal counts for each of these regions were used to calculate the ischemic-to-nonischemic ratios.
Gamma Well Counting of Tissue Activity
The skeletal muscles in hindlimb were excised and divided into proximal and distal sections according to the location of the ligature in the ischemic limb or to the anatomic landmarks in the contralateral limb. The tissue samples were weighed, and the tissue radioactivity was measured with a gamma well counter (Cobra Packard). Tissue samples were counted for 99mTc, and activity was corrected for background, decay time, and tissue weight.
Cross-sectional tissue blocks (2 mm thick) were taken from the middle of the ischemic lower hindlimb below the site of occlusion for immunohistochemical analyses. Specimens were embedded in TissueTec (Sakura) and snap-frozen in −150°C methylbutane. Frozen sections (5 μm) were fixed with acetone, incubated with biotinylated Griffonia Bandeiraea Simplicifolia Isolectin I, and visualized with the use of avidin-biotin horseradish peroxidase visualization systems (Vectastain ABC Kit Elite, Vector Laboratories). Capillary densities were calculated quantitatively for extent (percent area) of positive staining in 4 randomly chosen high-powered fields (×200 magnification) within the ischemic area with the use of computer algorithms developed and validated in our laboratory, as previously reported.6
Frozen sections (5 μm) taken from the ischemic and nonischemic lower hindlimb were fixed with acetone and incubated with a cy-3 fluorescent endothelial cell marker (1:100, CD31) and costained with a fluorescein-labeled analogue of NC100692 (1:100). The images were captured with the use of a fluorescence microscope (Olympus America Inc) with a ×36 objective and processed with commercial software (Adobe Photoshop, Adobe).
The Student t test was used to compare 2 groups. One-way ANOVA was used to compare multiple parameters. A value of P<0.05 was considered significant.
Figure 1 shows murine microvascular endothelial cells stained with a fluorescein-labeled homologue of NC100692 (RGD). The specificity of NC100692 for endothelial cells was demonstrated by uptake of only the targeted RGD peptide in focal contacts of the endothelial cells and the absence of focal staining with a control non-RGD peptide (GDF).
Pinhole Planar Imaging
Dynamic pinhole planar imaging immediately after injection of NC100692 demonstrated rapid clearance of the radiotracer from the blood with primary clearance through the kidneys (data not shown). Delayed 15-minute static images acquired 75 minutes after the intravenous injections were of excellent quality and allowed for quantitative analysis of radiotracer retention in the ischemic hindlimb. Figure 2A shows representative planar pinhole images recorded at 75 minutes after injection of NC100692 in control mice and mice at 1, 3, 7, and 14 days after femoral ligation. Differences in the image-derived ratios of radiotracer retention (ischemic to nonischemic) are shown in Figure 2B. A significant (P<0.05) increase in relative radiotracer retention (“hot spot”) was observed at day 3 (1.5±0.7) and day 7 (1.6±0.8) relative to control (0.9±0.1). The retention ratio of NC100692 was not increased at days 1 and 14 compared with control (Figure 2B).
Gamma Well Counting of Tissue Activity
NC100692 activity was measured in the skeletal muscles of the excised hindlimbs immediately after each imaging session. The ischemic-to-nonischemic activity ratios derived by gamma well counting are presented in Figure 2C. A significant (P<0.05) increase in NC100692 activity in the muscles distal to the femoral occlusion relative to the contralateral control hindlimb was observed on day 3 (2.3±1.1) and day 7 (2.7±0.7) relative to control mice (0.9±0.1). Although these ratios were higher than those observed by in vivo imaging, they were highly correlated with the imaging results (R2=0.73). The ischemic-to-nonischemic activity ratios were not increased in the muscle proximal to the occlusion.
To verify the presence of angiogenesis within the distal ischemic hindlimb, we evaluated capillary and arteriolar density with lectin staining. As illustrated in Figure 3A, there was an increase in capillary density in the distal ischemic hindlimb by lectin staining as early as 3 days after occlusion. Quantitative analysis of the lectin staining is shown in Figure 3B.
To confirm the specificity of the binding of our tracer targeted at the αvβ3 integrin to the angiogenic endothelial cells, we evaluated the costaining of a fluorescent analogue of NC100692 (RGD) in tissue specimens with a well-established fluorescent marker of endothelial cells (CD31). The representative images of the specimens taken from ischemic and contralateral nonischemic hindlimb at 7 days after femoral artery ligation are shown in Figure 4.
In the control hind limb, there were many CD31 positively stained endothelial cells within capillaries and small arterioles, although there was no RGD staining. In contrast, there was significant RGD staining of endothelial cells within the ischemic hindlimb, which colocalized with CD31 staining. Interestingly, not all endothelial cells colocalized with our targeted tracer. These endothelial cells within the ischemic tissue that did not stain with the αvβ3-targeted fluorescent analogue were probably more mature and quiescent.
This study demonstrates the temporal changes in ischemia-induced angiogenesis in the murine hindlimb with the use of high-resolution gamma camera imaging of a 99mTc-labeled chelate-peptide conjugate (NC100692) targeted at the αvβ3 integrin. To noninvasively evaluate the time course of αvβ3 integrin expression in the ischemic hindlimb, we imaged the animals at different time points after femoral occlusion. Increased focal NC100692 retention was seen as early as 3 days after onset of ischemia and peaked at 7 days after occlusion. At the later time points, NC100692 activity within the hindlimb gradually decreased to background levels. A fluorescent analogue of NC100692 was used to confirm localization of the targeted tracer in cultured endothelial cells, and endothelial cell specificity was further confirmed on tissue sections with the use of dual immunofluorescent staining of endothelium and the fluorescent compound.
To validate our noninvasive imaging approach, gamma well counting of tissue activity was used as a “gold standard” in all animals. We found a strong positive correlation between relative retention of the αvβ3-targeted radiotracer assessed with the use of planar imaging and the gamma well counting results, although the analysis of in vivo NC100692 planar pinhole images tended to underestimate the magnitude of the relative increases in NC100692 retention within the ischemic hindlimb. This difference in magnitude may be due to attenuation and partial volume errors and could be minimized by use of a higher-resolution 3D micro-SPECT-CT imaging system instead of the planar imaging used in the present study. The future application of this newer hybrid imaging technology could allow for correction of these errors and better absolute in vivo noninvasive quantification of angiogenesis.
Our observations with regard to the relative increased expression or possible activation of αvβ3 integrin early after ischemic injury are in agreement with data obtained by other investigators using invasive techniques. Sun et al11 recently demonstrated temporal changes in β3 expression in rats after permanent coronary occlusion. The β3 expression in their model of ischemic injury in the heart increased by day 3, peaked at day 7, and gradually declined thereafter. We have observed a similar temporal increase in activation of the αvβ3 integrin in the dog model of myocardial injury using an alternative peptidomimetic, which also targeted the αvβ3 integrin.6 Additional studies with this alternative radiolabeled agent targeted at the αvβ3 integrin have demonstrated preferential binding to the activated conformation of the αvβ3 integrin.8 Other investigators have reported a significant increase in both relative blood flow and hemoglobin oxygen saturation in a similar model of ischemic hindlimb at days 3 and 7, reaching a semiplateau on day 14 after femoral occlusion.12 These physiological changes would be expected to parallel changes in biological markers of angiogenesis like the αvβ3 integrin. Unfortunately, there is no antibody for the αvβ3 integrin that works in the mouse, and therefore it is difficult for us to confirm the specific increase in expression of the αvβ3 integrin in our model. The specificity of NC100692 for the αvβ3 integrin has been previously demonstrated in vitro.7 In the present study we confirmed the binding of a fluorescent analogue of the αvβ3-targeted compound to microvascular endothelial cells and demonstrated colocalization of the fluorescent analogue with a subset of endothelial cells in tissue sections. However, we cannot exclude the potential binding of NC100692 to other inflammatory cells or smooth muscle cells in the ischemic limb. Additional preclinical studies will be required to address the potential binding of NC100692 to these other cell types involved with ischemia-induced angiogenesis.
The targeted imaging of angiogenesis with NC100692 provides a novel, highly sensitive noninvasive approach for direct evaluation of the αvβ3 integrin, a biological marker known to modulate angiogenesis. This targeted biological approach offers advantages over other more standard measures of the physiological consequences of angiogenesis-like changes in tissue perfusion. Moreover, because of the short half-life of 99mTc, the present approach offers an ability to serially evaluate the angiogenic process in animals without the need to euthanize them at distinct time points. The focal retention and favorable clearance kinetics of NC100692 suggest the potential for future clinical in vivo imaging of angiogenesis in patients with peripheral vascular disease.
This research was supported by grants from the National Institutes of Health, National Heart, Lung, and Blood Institute (HL65662, to Dr Sinusas); American Heart Association (0430038N, to Dr Sadeghi); and Amersham Health.
Dr Mendizabal was a full-time employee of GE Healthcare at the time this study was conducted.
↵*The first 2 authors contributed equally to this work.
Fam NP, Verma S, Kutryk M, Stewart DJ. Clinician guide to angiogenesis. Circulation. 2003; 108: 2613–2618.
Morrison MS, Davis J, Ricketts S-A, Cuthbertson A, Mendizabal MV. Monitoring of tumour response to therapy with a novel angiogenesis imaging agent. Mol Imaging. 2003; 2: 272.
Sadeghi MM, Krassilnikova S, Zhang J, Gharaei AA, Fassaei HR, Esmailzadeh L, Kooshkabadi A, Edwards S, Yalamanchili P, Harris TD, Sinusas AJ, Zaret BL, Bender JR. Detection of injury-induced vascular remodeling by targeting activated αvβ3 integrin in vivo. Circulation. 2004; 110: 84–90.
Silvestre JS, Mallat Z, Duriez M, Tamarat R, Bureau MF, Scherman D, Duverger N, Branellec D, Tedgui A, Levy BI. Antiangiogenic effect of interleukin-10 in ischemia-induced angiogenesis in mice hindlimb. Circ Res. 2000; 87: 448–452.
Sun M, Opavsky MA, Stewart DJ, Rabinovitch M, Dawood F, Wen WH, Liu PP. Temporal response and localization of integrins beta1 and beta3 in the heart after myocardial infarction: regulation by cytokines. Circulation. 2003; 107: 1046–1052.