Background
The formation of new capillaries from existing micro vessels occurs as a natural healing process following myocardial infarction (MI) [
1]. This process, called angiogenesis, is triggered by ischemia and results in partial restoration of blood perfusion in the ischemic zone. Importantly, the extent of angiogenesis is associated with post-infarct remodelling and has positive implications for the prognosis of MI patients [
2]. Several therapy approaches, aimed at stimulating angiogenesis in the infarcted area, including gene therapy, intramyocardial administration of pro-angiogenic factors, administration of bone marrow-derived stem cells, and transmyocardial revascularization have been tested [
3-
12]. To date, results in animal models have been encouraging [
3-
5]. Nonetheless, the clinical translation of these therapies has proven to be difficult, and the clinical benefit for MI patients has been disappointing and controversial [
6-
12].
In order to improve treatment options, assessing the efficacy of pro-angiogenic therapies and monitoring the temporal development of angiogenesis, non-invasive methods (e.g. molecular imaging) are warranted. A sensitive and specific molecular imaging marker for angiogenesis, however, is still missing.
Angiogenic vessels demonstrate upregulation of CD13, a membrane-bound aminopeptidase, on activated endothelial cells [
13,
14]. The cyclic tripeptide Asn-Gly-Arg (cNGR) homes to CD13 expressed on endothelial cells of angiogenic tumour vasculature and angiogenic vessels in the infarcted area and border zones of the myocardium [
14-
19]. Moreover, evidence was obtained that cNGR does not bind to CD13-positive macrophages in the infarcted myocardium and infarct border zone. This finding was ascribed to cNGR, possibly targeting a subset of post-translationally modified CD13 that might be specific to smaller and perhaps newly formed vessels [
17]. Furthermore, using polymerase chain reaction, peak expression of CD13 was shown 7 days after MI in a mouse myocardial infarction model [
15]. Non-invasive imaging of CD13 in mice has further been demonstrated with cNGR-labelled paramagnetic quantum dots for molecular magnetic resonance imaging (MRI) [
20]. Although angiogenic activity could selectively and non-invasively be depicted in the infarcted heart, molecular imaging with MRI requires high contrast agent dosage, which hampers translation to clinical trials.
The arginine-glycine-aspartic-acid (RGD) motif is known to label angiogenic active endothelium through binding to α
vβ
3 integrin [
21]. In several pre-clinical studies in different neovascularisation models, RGD conjugated agents have been employed [
2,
22]. However, competition studies in tumour angiogenesis with the NGR and RGD motif demonstrated a threefold higher target homing ratio (tumour/control organ) for NGR than that for RGD [
16].
In this study, we designed and developed a CD13 molecular imaging probe that allowed sensitive visualisation using micro-SPECT imaging. In order to target CD13, we used the cNGR tripeptide, having a tenfold higher targeting efficacy than the linear entity [
23], to design a non-invasive angiogenesis imaging probe. Thereto, the cNGR tripeptide was conjugated with diethylene triamine pentaacetic acid (DTPA), radiolabeled with
111In and evaluated in a mouse model of myocardial infarction in combination with
99mTc-sestamibi in dual-isotope micro-SPECT imaging. In order to determine binding specificity of our newly designed probe, we employed the validated subcutaneous Matrigel plug assay as a model for
in vivo angiogenesis.
Discussion
Angiogenesis is associated with post-infarct remodelling and has important implications for prognosis following MI [
2]. While previous research has mainly focused on developing pro-angiogenic therapies to stimulate angiogenesis in the infarcted myocardium [
28,
29], a clinically applicable method for the evaluation of these therapies is still lacking. Most clinical studies thus far have relied on indirect evidence of myocardial angiogenesis. For example, changes in stress-induced ischemia and myocardial perfusion were assessed by treadmill exercise testing and dual-isotope gated SPECT thallium-201/
99mTc-sestamibi perfusion SPECT imaging at baseline, day 29, day 57 and day 180 after administration of the pro-angiogenic recombinant-fibroblast growth factor 2 (rFGF2) in a clinical study [
9]. In another study, using a single intracoronary infusion of rFGF2, patients experienced increased quality of life (assessed by Seattle Angina Questionnaire) and exercise tolerance (assessed by treadmill exercise testing). Moreover, in this study, MRI demonstrated increased regional wall thickening and a reduction in the extent of the ischemic area compared to baseline [
30].
A clinical trial was performed in which patients with exertional angina were treated with the pro-angiogenic recombinant human vascular endothelial growth factor (rhVEGF) [
31]. In this trial, no improvement in exercise treadmill test time or myocardial perfusion, as assessed by rest-stress thallium-201/
99mTc-sestamibi (gated) SPECT scanning protocol, was detected. However, patients experienced a significant improvement in angina class at day 120 after the first administration of rhVEGF [
31]. Furthermore, in a small phase 1 clinical study using VEGF gene transfer, all patients had significant reduction in angina and objective evidence of reduced ischemia was documented using dobutamine
99mTc-sestamibi SPECT imaging. Post-operative left ventricular ejection fraction (LVEF) was either unchanged or improved. Moreover, the Rentrop scoring system, indicating the coronary collateral filling as assessed by coronary angiography, improved in all patients [
32].
The pro-angiogenic effect of transmyocardial laser revascularisation (TMLR) was assessed in a small clinical study (seven patients) by Rimoldi et al. While angina scores (Canadian Cardiovascular Society) decreased significantly, exercise tolerance (treadmill testing), 12-min walk distance and the LVEF (assessed by radionuclide ventriculography) remained unchanged. Moreover, despite subjective improvement in some patients, myocardial blood flow (as assessed by
15O-labelled water) during dobutamine in lasered segments failed to show any significant increase [
33].
Clearly, improved myocardial perfusion or function can be used as indirect evidence for angiogenesis in the infarcted myocardium. Reduced angina scores, increased treadmill exercise times and improved myocardial perfusion as measured by angiography, SPECT, PET and MRI all seem to indicate recovery [
34]. However, despite the ability of these methods to indicate recovery, it often takes several months before any improvement can be detected. Therefore, these methods, used in current MI patient care, lack the sensitivity to monitor, or predict, pro-angiogenic treatment. To make direct angiogenesis imaging available for MI patient care, sensitive and specific non-invasive direct angiogenesis imaging probes are required.
To our knowledge, this is the first time CD13 expression was visualised with a nuclear cNGR probe in a mouse model of MI using dual-isotope micro-SPECT imaging. Simultaneous perfusion-imaging enabled us to visualise the angiogenesis process in the border zone of the infarct and the infarcted tissue itself. According to previous studies from our lab, angiogenesis is expected to occur in the infarct area and border zone of the infarct [
15,
20]. Additionally, in co-localisation experiments using the fluorescently labelled endothelial cell marker anti-CD31-PE, the fluorescently labelled angiogenic active endothelial cell marker anti-CD105-FITC and fluorescently labelled cNGR conjugates, evidence for preferential binding of cNGR conjugates to angiogenic active endothelial cells of capillaries and larger vessels with an inner diameter up to 15 μm was provided [
15,
20]. Moreover, despite the abundant presence of CD13-positive macrophages, fluorescently labelled cNGR did not bind to anti-CD13-FITC labelled macrophages [
15]. In the current study, we used the exact same animal model and the same cNGR moiety, assuring identical experimental conditions. Extrapolating from these data, we therefore consider binding of our
111In-DTPA-cNGR imaging probe to be specific for angiogenic endothelium.
Specific binding of our probe is further emphasised by enhanced binding of 111In-DTPA-cNGR to angiogenic active endothelium in a specific model for angiogenesis. Unfortunately (hFGF2-supplemented) Matrigels, implanted in the flanks for optimal stimulation of angiogenesis, were unrecognisable on in vivo SPECT scans as high renal uptake hampered image analysis. Therefore, we explanted the Matrigels and performed ex vivo SPECT scans that indicated higher uptake of 111In-DTPA-cNGR in hFGF2-supplemented Matrigels compared with control Matrigels. Moreover, the uptake of 111In-DTPA-cNGR in hFGF2-supplemented Matrigels was higher compared with that in the control Matrigels as indicated by gamma counting. hFGF2-supplemented Matrigels displayed significantly higher uptake than control muscle tissue, indicating active angiogenesis.
To use this imaging probe in a clinical setting, the probe should display a high specific target uptake and fast clearance from non-specific organs. The biodistribution characteristics of our probe were studied 2 h post-injection and demonstrated a rapid clearance in the urine which is a favourable clearance pathway. This is in contrast with results for the hydrophilic RGD probe, mainly studied for tumour angiogenesis imaging, that demonstrates unfavourable binding to kidneys, spleen, lungs, liver, and gastrointestinal system thereby hampering the image quality [
35].
We simultaneously imaged and quantified cardiac uptake of 99mTc and 111In in the AHA 17 segment model. The isotope uptake per segment was quantified using isotope specific CFs, assessed in phantom studies. Furthermore, isotope uptake was expressed in the clinically used SUV parameter, which normalises for injected dose and body weight. Although normalisation for body weight might be considered irrelevant for a rapidly excreted tracer that does not reach a uniform distribution throughout the body, we believe that the body weight bias in our quantification is limited due to the fact that all the animals used in this study were of similar age and body weight.
We found target-specific binding of our new angiogenesis imaging probe as evidenced by the significantly enhanced SUVs in six segments in the infarcted myocardium and the infarct border zone. To even further improve dual-isotope micro-SPECT imaging of perfusion defects and mirroring
111In-DTPA-cNGR uptake in the hearts of the MI mice, we are currently pursuing strategies to further enhance probe-target interaction through multivalent binding which may result in higher image quality. The proof of principle of this approach has already been demonstrated in studies from Li et al. and Dijkgraaf et al., in which tumour uptake increased with increasing RGD valency [
35,
36].
Despite displaying enhanced, specific uptake in the infarcted myocardium and infarct border zone, we observed several remarkable and so far inexplicable features of our probe. We found the overall uptake to be higher in every section of the infarcted hearts compared with that in the healthy control hearts. This may be due to enhanced post-infarct overall stress in the heart, but we cannot confirm such a finding at this moment. Moreover, kidney uptake was significantly higher in the healthy control mice compared to that in the MI mice. Although we currently have no direct link between the infarct and these findings, these discoveries we made using SPECT scans were overlooked in earlier studies employing two-photon microscopy and MRI and may be of considerable importance for future studies employing the cNGR molecule as a molecular angiogenesis imaging probe.
In this study, we opted for 111In due to the fact that pre-clinical SPECT scanners outclass pre-clinical PET scanner in terms of spatial resolution. Of course, the use of 111In in a clinical situation may be less favourable due to its long physical half-life of 2.8 days and resolution limitations. While pre-clinical SPECT scanners outclass pre-clinical PET scanners in terms of spatial resolution, in the clinical setting, this is reversed. Therefore, angiogenesis imaging via the cNGR tripeptide in MI patients would benefit from the use of a clinical PET system. Future research should therefore focus on the development of a high-resolution cNGR-based PET imaging probe, for example by replacing 111In by the PET isotope 68Ga.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
GH participated in the conception and study design, acquisition of data, data analysis, interpretation and writing. MdSH participated in the conception and study design, data interpretation, and writing. ID and MB participated in the data interpretation and drafting of the manuscript and revised it critically for important intellectual content. KD participated in the drafting of the manuscript and revised it critically for important intellectual content. RW carried out the data analysis and critically revised the manuscript for important intellectual content. IP and NvDa critically revised the manuscript for important intellectual content. TH, MP, and FM critically revised the manuscript for important intellectual content and gave final approval of the version to be published. All authors read and approved the final manuscript.