*Article* **Improved Detection of Molecular Markers of Atherosclerotic Plaques Using Sub-Millimeter PET Imaging**

**Jessica Bridoux 1, Sara Neyt 2, Pieterjan Debie 1, Benedicte Descamps 3, Nick Devoogdt 1, Frederik Cleeren 4, Guy Bormans 4, Alexis Broisat 5, Vicky Caveliers 1,6, Catarina Xavier 1, Christian Vanhove <sup>3</sup> and Sophie Hernot 1,\***


Academic Editors: Anne Roivainen and Xiang-Guo Li Received: 30 March 2020; Accepted: 13 April 2020; Published: 16 April 2020

**Abstract:** Since atherosclerotic plaques are small and sparse, their non-invasive detection via PET imaging requires both highly specific radiotracers as well as imaging systems with high sensitivity and resolution. This study aimed to assess the targeting and biodistribution of a novel fluorine-18 anti-VCAM-1 Nanobody (Nb), and to investigate whether sub-millimetre resolution PET imaging could improve detectability of plaques in mice. The anti-VCAM-1 Nb functionalised with the novel restrained complexing agent (RESCA) chelator was labelled with [18F]AlF with a high radiochemical yield (>75%) and radiochemical purity (>99%). Subsequently, [18F]AlF(RESCA)-cAbVCAM1-5 was injected in ApoE−/<sup>−</sup> mice, or co-injected with excess of unlabelled Nb (control group). Mice were imaged sequentially using a cross-over design on two different commercially available PET/CT systems and finally sacrificed for ex vivo analysis. Both the PET/CT images and ex vivo data showed specific uptake of [18F]AlF(RESCA)-cAbVCAM1-5 in atherosclerotic lesions. Non-specific bone uptake was also noticeable, most probably due to in vivo defluorination. Image analysis yielded higher target-to-heart and target-to-brain ratios with the β-CUBE (MOLECUBES) PET scanner, demonstrating that preclinical detection of atherosclerotic lesions could be improved using the latest PET technology.

**Keywords:** vulnerable plaque; molecular imaging; PET imaging; nanobody; single-domain antibody; sub-millimetre resolution; AlF-radiolabelling

#### **1. Introduction**

Atherosclerosis is the progressive narrowing of arteries caused by the accumulation of lipids and fibrous elements in the artery walls. Although lesion growth will lead to progressive blood vessel occlusion, a large proportion of patients show no sign of disease until the sudden rupture of so-called vulnerable plaques. Rupture is usually associated with thrombosis, causing myocardial infarctions, strokes or peripheral vascular disease [1]. Together, those severe clinical complications claim over 15 million lives every year, making cardiovascular diseases the leading cause of death worldwide [2]. The joint ESC Guidelines suggested that early diagnosis of high-risk patients could be equally effective as preventing new cases, leading to potential cost-savings, and consequently encouraged research regarding risk assessment by non- or minimally invasive imaging techniques [3]. Techniques such as multi-detector CT [4,5], intravascular ultrasound [6,7], MRI [8,9], or Optical Coherence Tomography [10] are extensively being investigated for the assessment of morphological or structural characteristics of atherosclerotic lesions. Among the molecular imaging modalities, which can reveal specific biological aspects of atherosclerotic plaques, positron emission tomography/computed tomography (PET/CT) is one of the preferred techniques in clinic [11]. Although PET/CT is a sensitive and quantitative technique, most of the commercially available pre-clinical PET scanners do not meet the necessary sensitivity and spatial resolution to fully support clinical translation of new promising tracers [12]. Recently a novel PET scanner (β-CUBE, Molecubes, Ghent, Belgium) became commercially available that uses monolithic scintillation detectors to obtain sub-mm spatial resolution in combination with high sensitivity [13], which might improve plaque detection in mice.

In order to visualise the recruitment of inflammatory cells in atherosclerotic plaques with PET imaging, we used a Nanobody (Nb)-based tracer (cAbVCAM1-5) targeting the vascular cell adhesion molecule-1 (VCAM-1) [14,15]. Nbs are small antigen-binding fragments derived from heavy-chain-only antibodies and proved to have ideal characteristics for PET imaging [16,17]. Furthermore, the biological half-life of Nbs matches the half-life of fluorine-18 (18F) (109.8 min), the most commonly used positron-emitting isotope because of its favourable nuclear decay characteristics. 18F-labelling of heat-sensitive biomolecules is commonly performed via prosthetic groups. However, this time-consuming process often has low efficiency. Herein, we overcame the previous issues by functionalising the Nb with the novel restrained complexing agent (RESCA) developed by Cleeren et al. [18], allowing fast and simple 18F-labelling via the Al18F-method [19,20].

In this study, the cAbVCAM1-5 Nb, was labelled via Al18F(RESCA) chemistry, and evaluated as a tracer to image atherosclerosis plaques in Apolipoprotein E-deficient (ApoE−/−) mice. In addition, we investigated whether the imaging could be improved using the latest β-CUBE PET technology.

#### **2. Results**

#### *2.1. Conjugation with RESCA and Radiolabelling of the Nb*

The produced cAbVCAM1-5 Nb was randomly modified through conjugation of tetrafluorophenyl TFP-RESCA (Figure 1) on its lysines for subsequent Al18F-labelling. Electrospray ionisation and quadrupole time-of-flight mass spectrometry (ESI-Q-ToF-MS) analysis revealed successful conjugation of the cAbVCAM1-5 Nb with RESCA. For cAbVCAM1-5-(RESCA)n, a mass of 14,658 + *n* × 419.5 Da was expected. Measured mass was obtained for *n* = 1 (15,076 ± 2) Da, *n* = 2 (15,495 ± 2) Da, *n* = 3 (15,913 ± 2) Da and *n* = 4 (16,331 ± 2) Da (Figure S1).

**Figure 1.** Structure of tetrafluorophenyl restrained complexing agent (TFP-RESCA).

Next, cAbVCAM1-5 randomly conjugated with RESCA was radiolabelled at room temperature (RT) with [18F]AlF with a 78 <sup>±</sup> 2% radiochemical yield (RCY). Separation of Nb from free [18F]AlF was performed through a desalting PD10 column which was eluted in 500 μL fractions. The two fractions containing most of the activity were combined and filtered, allowing to obtain a radiochemical purity (RCP) of 99% (Figure 2) and an apparent molar activity of 24.5 ± 3.1 GBq/μmol. The radiolabelling and purification procedures were completed in less than an hour. [18F]AlF(RESCA)-cAbVCAM1-5 Nb remained stable with a RCP of 96% (Figure S2A) over 3 h 30 min in injection buffer at RT, as well as in human serum at 37 ◦C over 1 h 30. At 2 h 30 min up to 6% defluorination was observed in human serum (Figure S2B).

**Figure 2.** Size Exclusion Chromatography (SEC) profile of [18F]AlF(RESCA)-cAbVCAM1-5 Nb before injection. Retention time (Rt) of [18F]AlF(RESCA)-cAbVCAM1-5 = 28.7 min (99%), free [18F]AlF and [ 18F]F-Rt = 35.3 min (1%).

#### *2.2. Imaging with the* β*-CUBE and LabPET8 Systems*

In vivo PET imaging showed excretion of the tracer via the kidneys and bladder. The cohort injected with the [18F]AlF(RESCA)-cAbVCAM1-5 Nb showed substantial signal in bone structures (Figure 3A, upper row). This signal was also observed in the control group (Figure 3A, lower row), where the [18F]AlF(RESCA)-cAbVCAM1-5 Nb was co-injected with excess of unlabelled cAbVCAM1-5 Nb, indicating the non-specific character of the uptake.

Accumulation of [18F]AlF(RESCA)-cAbVCAM1-5 Nb in the aortic arch of ApoE−/<sup>−</sup> mice was observed, which is the predominant site for atherosclerotic lesion formation in this model (Figure 3A, upper row). No signal was observed in the aortic arch of the control group (Figure 3A, lower row).

**Figure 3.** *Cont.*

**Figure 3.** (**A**) Representative PET/CT images of the same mouse obtained with the LabPET8 (left) or β-CUBE (right) imaging system, demonstrating specific targeting of atherosclerotic lesions in the aortic arch (Ao) of ApoE−/<sup>−</sup> mice injected with [18F]AlF(RESCA)-cAbVCAM1-5 Nb (upper row), while no uptake is seen at the level of the aortic arch of ApoE−/<sup>−</sup> mice co-injected with a 90-fold excess of unlabelled cAbVCAM1-5 Nb (blocking condition as control, unlabelled excess injected 15 min before injection of radiolabelled Nb) (lower row). Kidneys (K), bladder (Bl) and bone structures (Bs) are also visible on the images. Target-to-brain (T/B) (**B**) and target-to-heart (T/H) (**C**) ratios were calculated to compare the image quality between two commercially available preclinical PET scanners (β-CUBE and LabET8). The number of asterisks in the figures indicates the statistical significance (\* *P* < 0.05).

When comparing the imaging data obtained with two distinct preclinical PET devices in a crossover study, better image quality was achieved with the β-CUBE than with the LabPET8 (Figure 3A). In vivo image contrast was evaluated by calculating target-to-brain (T/B) and target-to-heart (T/H) ratios. In both cases, significantly higher values were obtained with the β-CUBE than with the LabPET8 (T/B: 3.88 ± 0.88 vs. 2.57 ± 0.54, *p* < 0.05; T/H: 1.75 ± 0.30 vs. 1.40 ± 0.24, *p* < 0.05; respectively).
