Association between [68Ga]NODAGA-RGDyK uptake and dynamics of angiogenesis in a human cell-based 3D model

Radiolabeled RGD peptides targeting expression of αvβ3 integrin have been applied to in vivo imaging of angiogenesis. However, there is a need for more information on the quantitative relationships between RGD peptide uptake and the dynamics of angiogenesis. In this study, we sought to measure the binding of [68Ga]NODAGA-RGDyK to αvβ3 integrin in a human cell-based three-dimensional (3D) in vitro model of angiogenesis, and to compare the level of binding with the amount of angiogenesis. Experiments were conducted using a human cell-based 3D model of angiogenesis consisting of co-culture of human adipose stem cells (hASCs) and of human umbilical vein endothelial cells (HUVECs). Angiogenesis was induced with four concentrations (25%, 50%, 75%, and 100%) of growth factor cocktail resulting in a gradual increase in the density of the tubule network. Cultures were incubated with [68Ga]NODAGA-RGDyK for 90 min at 37 °C, and binding of radioactivity was measured by gamma counting and digital autoradiography. The results revealed that tracer binding increased gradually with neovasculature density. In comparison with vessels induced with a growth factor concentration of 25%, the uptake of [68Ga]NODAGA-RGDyK was higher at concentrations of 75% and 100%, and correlated with the amount of neovasculature, as determined by visual evaluation of histological staining. Uptake of [68Ga]NODAGA-RGDyK closely reflected the amount of angiogenesis in an in vitro 3D model of angiogenesis. These results support further evaluation of RGD-based approaches for targeted imaging of angiogenesis.


Introduction
Integrins are transmembrane glycoprotein receptors that mediate interactions between cells and their surroundings during normal development, and in response to physiological and pathophysiological signals. Integrin α v β 3 plays a key role in angiogenesis associated with various disease states, such as ischemic heart disease and during tumor growth [1,2]. Expression of integrin α v β 3 is generally low in endothelial cells but rises during angiogenesis [1], when the receptor also undergoes a conformational change into the active form [3].
Integrin specific ligands containing the tripeptide sequence arginine-glycine-aspartic acid (RGD) have been radiolabeled for non-invasive imaging of integrin expression using single-photon emission computed tomography (SPECT) or positron emission tomography (PET) [4]. Radiolabeled peptides that selectively bind to α v β 3 integrin have shown promise for non-invasive imaging of pathological conditions associated with angiogenesis, such as peripheral artery disease, ischemic myocardial injury, stroke and cancer [5][6][7][8]. In such studies, the uptake of RGD-based tracers has been associated with histological evidence of angiogenesis. However, some studies have suggested that uptake reflects other processes, such as inflammation [9,10] and fibrosis [11,12]. Thus, further evaluation of the relationship between the uptake of RGD-based tracer and dynamics of angiogenesis is needed to understand whether the activity of angiogenesis can be accurately quantified by imaging.
[ 68 Ga]NODAGA-RGDyK is a novel 68 Ga-labeled cyclic 1,4,7-triazacyclononane-1-glutaric acid-4,7-diacetic acidconjugated RGD peptide designed for PET imaging of α v β 3 integrin expression. It has favorable biokinetics, radiation dose, and safety profile [13,14]. In a previous study using [ 68 Ga]NODAGA-RGDyK, we detected an increase in myocardial α v β 3 integrin expression in ischemic myocardium, localized within irreversibly injured myocardium, in the absence of an increase in angiogenesis [12]. To further evaluate the ability of [ 68 Ga]NODAGA-RGDyK to measure angiogenic activity, we investigated whether binding of this agent reflects the amount of angiogenesis in a human cell-based 3D in vitro model of angiogenesis [15,16]. This model consisting of a co-culture of human umbilical cord vein endothelial cells (HUVECs) and human adipose stromal cells (hASCs) mimics dynamics of the development of normal human vasculature. After stimulation with growth factors, the cells produce a vascular-like tubule network with morphology of mature vessels within few days. The density of the vascular network can be reproducibly controlled by adjusting concentrations of growth factors.

Angiogenesis tissue model
We used a human cell-based 3D model of angiogenesis developed at the Finnish Centre for Alternative Methods (FICAM) [15,16]. When establishing the cell model, hASC cells were thawed, propagated for 9 days, and then seeded onto a plastic coverslip in 24-well plates (for receptor binding studies) or in 2-well chamber microscope slides (for autoradiography studies). On the same day, HUVECs that had been thawed and propagated for 3 days were plated on top of the hASC cell layer to form a co-culture. The next day, stimulation of the vascular-like tubule network formation was initiated by increasing the concentration of vascular endothelial growth factor (VEGF) (R&D Systems, Inc, Minneapolis, US) and fibroblast growth factor β (FGF-β) (R&D Systems, Inc, Minneapolis, US); induction medium was replaced with fresh medium after 3 days. To obtain a gradient in tubule network density, four different growth factor concentrations were used: 10 ng/ml VEGF and 1 ng/ ml FGF-β (defined as 100%), and 25%, 50%, and 75% of that maximum concentration. Co-culture of hASCs and HUVECs without induction of a tubule network (because inductive growth factors were not added) served as a control. Cell culture was terminated 6 days after the start of the tubule network stimulation.

In vitro autoradiography
Transport medium was removed from the chamber slides, and warm medium was added. Tissue cultures were incubated for 1 h at 37 °C; medium was removed; and [ 68 Ga] NODAGA-RGDyK was added. After incubation for 90 min at 37 °C, the tracer solution was removed, the chamber walls were removed, and the cells were washed twice with PBS. Slides were air dried for 10 min and exposed on a phosphorimaging plate (BAS-TR2025; Fuji Photo Film, Tokyo, Japan) for 1 h. The distribution of radioactivity on the plate was visualized and quantified using a Fluorescent Image Analyzer (Fujifilm FLA-5100, Fuji Photo Film) and the AIDA Image analyzer software v.4.55 (Raytest Isotopenmessgeräte GmbH, Straubenhardt, Germany). Autoradiography uptake was expressed as photostimulated luminescence per square millimeter (PSL/mm 2 ), and the vessel-to-base cell layer ratio was calculated by dividing the uptake of the vessels by the uptake of the base cell layer consisting of hASCs and HUVECs without growth factors.

Immunohistochemistry
After measurements with gamma counter or autoradiography, the tissues were fixed in 70% ethanol for 15 min and stained with antibodies against von Willebrand factor (vWF) and collagen IV. Cells were permeabilized with Triton X-100 for 15 min, and nonspecific binding sites were blocked with 10% BSA for 30 min. Primary antibodies (anti-von Willebrand factor, dilution 1:100; anti-collagen IV, dilution 1:500; Sigma-Aldrich, Saint Louis, Missouri, US) were applied overnight. After washing steps, secondary antibodies (tetramethylrhodamine [TRITC]-conjugated anti-rabbit for von Willebrand factor and fluorescein isothiocyanate [FITC]-conjugated anti-mouse for collagen IV; Sigma-Aldrich, Saint Louis, Missouri, US) were applied for 45 min. Secondary antibodies were washed away, and PBS was left in the wells. Fluorescence imaging of stained tubular structures was performed using a Cell-IQ automated image analysis system (CM-Technologies, Tampere, Finland). The degree of angiogenesis in each image was graded manually; the scoring criterion was the intensity of the vascular-like tubular network (tubule length and branching). The highest score was 5 (dense, highly branched tubular network), and the lowest score was 0 (no or minor tubules). In order to detect α v β 3 integrin, tissues were stained with antiintegrin α v β 3 antibodies (dilution 1:100, Sigma-Aldrich, Saint Louis, Missouri, US) and Alexa Fluor 488 conjugated anti-mouse secondary antibody.

Statistical analyses
SPSS Statistics software v.25 (IBM, Armonk, NY, USA) was used for statistical analyses. Unpaired Student's t-test was used for comparisons between two groups, and comparisons between multiple groups were performed using ANOVA with Tukey's post hoc test. Spearman's rank test (r s ) was used to analyze correlation. p < 0.05 was considered statistically significant.

In vitro tracer binding
We performed a tracer binding study in a human cell-derived angiogenesis tissue model using four different concentrations of angiogenesis-promoting growth factors, which yielded vascular-like tubule networks of four different densities. The density of tubules gradually increased with growth factor concentration (Fig. 1a). Relative to control cells that were not treated with growth factors, uptake of [ 68 Ga]NODAGA-RGDyK was 1.6-fold higher at 25% growth factor concentration, 2.1-fold higher at 50%, 2.4-fold higher at 75%, and 2.5-fold higher at 100%. Relative to the 25% concentration, uptake of [ 68 Ga]NODAGA-RGDyK was significantly higher at 75% (p = 0.0003) and 100% (p = 0.0002) (Fig. 1b). [ 68 Ga] NODAGA-RGDyK uptake also correlated with the density of vascular-like tubule network, as determined by visual examination of histological staining (Fig. 1c).

Autoradiography
Representative autoradiographs of [ 68 Ga]NODAGA-RGDyK uptake in vessels with growth factor concentrations of 75% and 100% are shown in Fig. 2. Uptake was localized in the regions of angiogenesis, indicated by tubule structures positive for vWF and collagen IV, and was significantly higher at 100% vs. 75% growth factor concentration (vessel-to-base cell layer ratio: 3.3 ± 0.09 vs. 2.5 ± 0.63; p = 0.009).

Discussion
In this study, using an in vitro 3D model of angiogenesis, we found that [ 68 Ga]NODAGA-RGDyK binding was proportional to the amount of angiogenesis. Our experiments were performed in a human cell-based angiogenesis tissue model consisting of a co-culture of HUVECs and human adipose stromal cells (hASCs), which generate a reproducible, extensive, mature vascular-like tubule network after stimulation with a growth factor cocktail [15,16]. This is the first study of an RGD targeting tracer in a tissue model that mimics post-infarct angiogenesis. The density of the tubule network increased with the concentration of growth factors. Vessels in this model exhibit mature properties and become surrounded by pericytes [16,18]. Pericyte coverage is a characteristic of maturation of newly formed vessels in the myocardium after infarction [2], indicating that the model we used accurately represents in vivo angiogenesis. α v β 3 integrin also seems to be involved in the maturation process, as evidenced by the fact that anti-α v β 3 integrin therapy decreases pericyte coverage in tumor vessels [19]. Our results lend support to the concept that [ 68 Ga]NODAGA-RGDyK can be used to quantitatively assess activity of angiogenesis, which is of particular importance for predicting and monitoring therapy responses.
Our results are in line with previous studies showing that the uptake of integrin-targeting RGD peptides is associated with histological evidence of newly formed blood vessels [6], coincides with hypoxia [20], and can be modified by therapy inducing [21,22] or inhibiting angiogenesis [23]. However, the limited number of time points over prolonged period (for example from 3 days to 3 months after an ischemic insult) does not permit detailed evaluation of association between tracer uptake and ongoing angiogenesis. Furthermore, in addition to mediating angiogenesis, α v β 3 integrin also regulates macrophage inflammatory responses and myofibroblast differentiation [24,25]. Therefore, the uptake of tracers targeting α v β 3 integrin may not merely represent angiogenesis, but also other events during the repair process after ischemic myocardial injury. Because multiple cell types express α v β 3 integrin, in vivo studies measuring RGD uptake in the context of angiogenesis have yielded conflicting results, as angiogenesis does not always co-localize with integrin expression [9][10][11][12]. Studied RGD-based peptides also have different selectivity for different types of integrins [26]. Our in vitro 3D model of angiogenesis provides a more controlled and isolated environment for newly formed vessels that lacks most of the other α v β 3 integrin binding cell types, such as myofibroblasts and macrophages, which are present in in vivo models of myocardial ischemia and infarction. This enables specific evaluation of Uptake of [ 68 Ga]NODAGA-RGDyK in vessels was elevated in the presence of higher growth factor concentrations and correlated with the abundance of the vessels. Due to problems with tissue attachment on slides at lower concentrations of growth factors, we performed autoradiography analysis only on the slides treated with concentrations of 75% and 100%; this represents a limitation of the study. However, we were able to detect a difference in [ 68 Ga]NODAGA-RGDyK binding between the two concentrations measured. Resolution of autoradiography did not allow detailed localization of the tracer signal with individual vessel structures. For this purpose, a different probe including an optical label could be used in experimental setting instead of a [ 68 Ga]NODAGA-RGDyK available for clinical use. Finally, detection of angiogenesis with RGD based tracers is dependent on integrin expression. Although α v β 3 integrin is typically expressed during angiogenesis, it may not always be present [27] and thus, there may not be tracer uptake under such circumstances.

Conclusions
[ 68 Ga]NODAGA-RGDyK correlated with the amount of angiogenesis in a human cell-based in vitro model of angiogenesis indicating that the tracer uptake reflects ongoing angiogenic activity. These results support further evaluation of RGD-based approaches for targeted imaging of angiogenesis in various disease states, such as ischemic heart disease and cancer. However, the possibility of nonspecific binding should be taken into account when performing in vivo imaging studies of angiogenesis.
Author contributions All authors contributed to the study conception and design. Material preparation, data collection, and analysis were performed by MG, OM, XGL, TT, and OH. All authors contributed to interpretation of data. The first draft of the manuscript was written by MG, and all authors commented on the drafts of the manuscript. All authors read and approved the final manuscript.

Conflict of interest
The authors declare that they have no conflicts of interest pertaining to this work.
Ethical approval hASCs and HUVECs were isolated from the human fat tissue and umbilical cords, respectively, as permitted by The Regional Ethics Committee of Tampere University Hospital's Responsibility Area in accordance with the International Conference on Harmonization (ICH) GCP guidelines (CPMP/ICH/135/95), the World Medical Association Declaration of Helsinki (1964), and subsequent amendments.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.