Autoradiography screening of potential positron emission tomography tracers for asymptomatic abdominal aortic aneurysms.

OBJECTIVE: The aetiology and early pathophysiological mechanisms of aortic aneurysm formation are still unknown and challenging to study in vivo. Positron emission tomography (PET) is a potentially valuable instrument for non-invasive in vivo pathophysiological studies. No specific tracer to identify the pathophysiological process of aneurysmal dilatation is yet available, however. The aim of this study was to explore if different PET tracers could be useful to image aneurysmal disease. METHODS AND
RESULTS: Human aneurysmal aortic tissue, collected during elective resection of abdominal aortic aneurysm (AAA) of asymptomatic patients, was investigated in vitro by means of autoradiography with [(68)Ga]CRP-binder targeting C-reactive protein, [(11)C]DAA1106 targeting translocator protein (18 kDa), [(11)C]D-deprenyl with unknown target receptor, [(11)C]deuterium-L-deprenyl targeting astrocytes, [(18)F]fluciclatide targeting integrin αVβ3, [(68)Ga]IMP461 and bi-specific antibody TF2 052107 targeting carcinoembryonic antigen, [(18)F]F-metomidate targeting mitochondrial cytochrome P-450 species in the adrenal cortex, and [(18)F]vorozole targeting aromatase. Of the investigated tracers, only [(18)F]fluciclatide exhibited specific binding, whereas the other PET tracers failed to show specific uptake in the investigated tissue and are probably not useful for the intended purpose.
CONCLUSION: It seems likely that αVβ3 integrin expression in AAA can be visualized with PET and that the αVβ3 selective tracer, [(18)F]fluciclatide, may be suitable for in vivo molecular imaging of asymptomatic AAA. Additional evaluation of [(18)F]fluciclatide and αVβ3 integrin expression in AAA will be performed in vitro as well as in vivo.

Introduction

The aetiology of degenerative abdominal aortic aneurysm (AAA) is mostly unknown. One of the observed features of the arterial wall in aneurysmal disease is degradation of connective tissue in the media layer. An increased proteolytic activity of matrix metalloproteinase, responsible for the degeneration of elastic lamellae and extracellular matrix proteins, has been demonstrated (1,2). Another observed characteristic is a chronic inflammation with large amounts of inflammatory cells, T- and B-lymphocytes, as well as macrophages (3). Furthermore, an increased medial neovascularization at the aneurysm rupture edge has been demonstrated. With angiogenesis as the predominant form of neovascularization in atherosclerosis it can be hypothesized that this process leads to a concomitant increase of αVβ3 integrin expression (4,5). Thus, there are multiple potential molecular targets or pathophysiological processes that, together with a suitable biomarker, could be of diagnostic or prognostic value in the management of AAA patients.

Methods for studies of the early pathophysiology in AAA in vivo are scarce. At the end stages of the disorder, biopsies may be obtained during surgery for histological investigations. A non-invasive diagnostic method that could differentiate between stable or metabolically active AAA at risk of rupture, based on knowledge at a molecular level, instead of size, would be of significant scientific and potential clinical value.

Positron emission tomography (PET) is a diagnostic imaging tool that provides a possibility for studies of pathophysiological mechanisms in vivo at the molecular level without interfering with that process. Together with computed tomography (CT), providing anatomical characteristics, the combined use of PET/CT has become a valuable instrument in various clinical settings, as well as a powerful research tool.

The aim of this study was to identify one or several useful PET tracers for the study of asymptomatic AAAs.

Methods

PET tracers

A number of tracers, originally intended for use in other disorders, were evaluated in vitro by means of autoradiography. Due to the infiltrate with inflammatory cells found in AAA, tracers with different characteristics of this field were tested: [68Ga]CRP-binder (CRPB), [11C]DAA1106 (DAA), and [11C]D-deprenyl (DDE). Since angiogenesis has been suggested as a possible cause of aneurysm formation [18F]fluciclatide was tested as well. Some tracers that were developed at the PET centre were tested on AAA tissue even though the likelihood of the intended receptors to be found in AAA tissue was very small: L-[11C]di-deuteriumdeprenyl (DED), TF2 052107 with [68Ga]IMP461 (TF2-IMP), [18F]fluorometomidate (FMTO), and [18F]vorozole (FVOZ).

All tracers were radiolabelled on site at the PET centre, Uppsala University Hospital, according to published methods. Some of the tracers were selected based on the potential of overexpression of certain receptors or enzymes in AAA tissue compared with normal aorta. Other tracers were tested merely based on availability.

[ (CRPB) (6) is a polypeptide with a phosphocholine group that has a moderate affinity to C-reactive protein (CRP) (Kd = 5 µM) (7). CRP is a protein that responds to inflammation and infections and can be detected in plasma (8). With the radiolabelled polypeptide targeting CRP it was hypothesized that the tissue responsible for the inflammation can be visualized in vivo. It might also be that this protein is expressed in inflammatory parts of AAA affected tissue.

[ is a ligand targeting the translocator protein (18 kDa), TSPO, also known as a peripheral benzodiazepine receptor (PBR), which is expressed on macrophages (9,10). It can be expected that macrophages are found in higher amounts in AAA tissue than in normal aorta. It has also been shown to be expressed in the brain in a variety of conditions: after stroke (11), in multiple sclerotic plaques (10,12), in dementia (13), and in refractory epilepsy (14). DAA has a high affinity to TSPO (Kd = 0.043 nM), which is an order of magnitude higher than the commonly used tracers [11C]PK11195 (15,16) and [11C]PBR28 (17).

[ is the inactive isomer of [11C]L-deprenyl and has an as yet unknown binding site, but has shown prospects in revealing different inflammatory states. Thus, it has been useful in the diagnosis of e.g. whiplash trauma and rheumatoid arthritis (18,19). Potentially, the unknown inflammatory-related binding site could also be expressed in AAA tissue.

[ (20) is an integrin receptor ligand, exhibiting high selectivity and affinity for αVβ3 (Kd ∼10 nM) (21). Integrin αVβ3, also known as vitronectin receptor, has been shown to be upregulated in angiogenesis (22,23). It is expressed on macrophages (24), where it seems to be a mere expression rather than being involved in their activation (25,26). Angiogenesis is a process expected to be upregulated in AAA (4) as well as in the vicinity of platelets (27), osteoclasts (28,29), endothelial cells (30), and malignant cells, such as in melanoma (31) and breast cancer (21).

L-[ is an irreversible inhibitor of monoamine oxidase type B (MAO-B) with high affinity and specificity (32,33). The tracer has been di-deuteriated in order to reduce the affinity towards the enzyme and minimize the blood flow dependency of the compound (34,35). In the diagnosis of epilepsy, this PET tracer, targeting astrocytes, has been of great value (36,37). Additional investigations of astrocytic invasion in patients with Alzheimer’s disease have been carried out with this tracer along with the amyloid-β-specific tracer [11C]PIB (38,39).

TF2-IMP is a bi-specific antibody coupled to a hapten. The TF2 052107 is a combined pretargeting antibody for carcinoembryonic antigen (CEA), which has a specific target site for the [68Ga]IMP461 (IMP) PET tracer (40). CEAs are produced during fetal development and are expressed in a variety of carcinomas, and can therefore be used as a biomarker both in tissue and in blood in those patients. Blood levels of CEA may also be elevated in other cancer types. The antibody is used as a pretarget of CEA and is subsequently visualized by the labelled hapten. The hapten IMP461(NOTA-D-Ala-D-Lys(HSG)-D-Tyr-D-Lys (HSG)-NH2) was obtained from Immunomedics, Inc. (Morris Plains, NJ, USA).

[ (FMTO) is an analogue to the more commonly used [11C]metomidate. It is useful in the visualization of adrenal cortical masses such as incidentalomas, adenomas, and primary and metastatic cortical carcinomas (41–43).

[. Vorozole, 6-[(4-chlorophenyl)(1,2,4-triazol-1-yl)methyl]-1-methylbenzotriazole, is a selective and potent non-steroidal aromatase enzyme inhibitor (44). [18F]Vorozole (45) is an analogue with similar affinity to aromatase (Kd = 0.21 nM) (46), an enzyme that converts androgen to estrogens. Aromatase inhibitors are used in the treatment of postmenopausal women with early-stage or advanced hormone-sensitive breast cancer. It has previously been labelled with the radionuclide 11C and used for visualization of aromatase distribution in the brain, especially the amygdala (46–48).

Experimental design

Human AAA tissue

The study was approved by the Regional Ethics Committee of Uppsala/Örebro (Dnr 2007/052). Three male patients operated on for asymptomatic AAA, aged 65, 68, and 74 years, were investigated. The AAA diameters were 52, 56, and 66 mm. Biopsies obtained from the anterior segment of the AAAs were 10 × 10 mm thick and consisted of the full thickness of the aortic vessel wall. They were embedded in Tissue-Tek® OCT™ Compound (Sakura Finetek, Alphen aan den Rijn, The Netherlands), immediately cryo-fixed on dry ice and isopentane in the operating room, and stored at –70°C.

Tracer syntheses

The tracers used were, apart from [18F]fluciclatide, synthesized according to standard procedures at the PET Centre, Uppsala University Hospital. [18F]Fluciclatide was synthesized according to an established method in a FASTlab and provided by Uppsala Imanet AB/GE Healthcare.

Frozen tissue autoradiography

Glass slides with sections of AAA tissue, 20–25 µm thick, were incubated in a buffer solution containing a known concentration of the radioligand, as well as with their blocking substances with the same concentration but including an excess of a blocking substance, for assessment of non-specific binding (Table I). After incubation, slides were washed in cold buffer followed by a brief wash in distilled water, dried, and exposed to phosphor imaging plates for at least three radionuclide half-lives. Non-blocked and blocked samples were exposed under the same plate. The plates were then scanned in a Phosphor Imager Model 400S using 100 µm pixel width (Molecular Dynamics, Sunnyvale, CA, USA), and the digital images were analysed using software ImageQuant 5.1 (Molecular Dynamics).

Table I.

The in vitro labelling conditions for the different tracers.

Abbreviation	Tracer	Molecular target	Concentration	Specific displacer	Buffer	Pre-incubation	Incubation	Wash (on ice)
CRPB	[68Ga]CRP-binder	C-reactive protein	0.01 MBq/mL, 0.205 nM	Phosphocholine, 10 µM	Tris-HCl 0.05 M, pH 7.4 + BSA 0.2% + MgCl2 5 mM + bacitracin 50 mg/L	10 min RT	60 min, 37°C	3 × 3 min
DAA	[11C]DAA1106	Translocator protein (18 kDa)	3 nM	PK11195, 1 µM	Tris-HCl 0.05 M, pH 7.4	–	40 min, RT	3 × 3 min
DDE	[11C]D-deprenyl	Unknown target receptor	10 nM	D-deprenyl, 1 µM	PBS + BSA, 1%, pH 7.4	–	40 min, 37°C	3 × 3 min
DED	[11C]-deuterium-L-deprenyl	Astrocytes	10 nM	L-deprenyl, 10 µM	10 min in PBS + BSA	10 min RT	40 min, 37°C	3 × 3 min
[18F]fluciclatide	[18F]fluciclatide	Integrin αvβ3	13 nM	Fluciclatide, 0.56 µM	Tris-HCl 0.05 M, pH 7.4	10 min RT	60 min, RT	3 × 3 min
TF2-IMP	[68Ga]IMP461 and bi-specific antibody TF2 052107	Carcinoembryonic antigen	0.25 MBq/mL	No antibody	Tris-HCl 0.05 M, pH 7.4, + BSA 0.2% + MgCl2 5 mM + bacitracin 50 mg/L	60 min at 37°C with TF2, wash 2×2 min wash on ice	60 min, 37°C	2 × 3 min
FMTO	[18F]F-metomidate	Mitochondrial cytochrome P-450 species in the adrenal cortex	0.007 MBq/mL	F-metomidate, 1 µM	Tris-HCl 0.05 M, pH 7.4	25 min, RT	40 min, RT	3 × 3 min
FVOZ	[18F]vorozole	Aromatase	0.1 MBq/mL, 2 nM	Vorozole, 1 µM	Tris-HCl 0.05 M, pH 7.4	10 min, RT	30 min, RT	3 × 2 min

BSA = bovine serum albumin; HCl = hydrochloric acid; M = molar; MBq = megabecquerel; nM = nanomolar; PBS = phosphate-buffered saline; RT = room temperature; Tris = tris(hydroxymethyl)aminomethane; µM = micromolar.

Image analysis

Autoradiograms were analysed using ImageQuant, and regions of interest (ROIs) were drawn manually on the digital images, delineating the whole segment of aortic tissue. Average pixel values were calculated, and the specific binding was determined as the difference between uptake with radioligand alone and that in the presence of an excess of blocking compound.

Results

Only [18F]fluciclatide displayed specific uptake (Figure 1), i.e. 88% of the uptake was blocked by co-incubation with excess of unlabelled fluciclatide, whereas the other tracers, CRP-binder, DAA, DDE, DED, TF2-IMP, FMTO, and FVOZ, did not display any specific uptake.

Figure 1.

Autoradiography with [18F]fluciclatide on AAA tissue from two patients. A: without additional blocking agent; and B: with additional blocking agent (unmarked fluciclatide).

Discussion

The natural course of AAA is to expand gradually and increase in diameter in order eventually to rupture, with a mortality rate as high as 90% (49). The only parameter to determine the time of intervention today is the diameter, usually obtained with ultrasound investigations. In order to intervene at an early phase of the pathology further knowledge on the pathophysiology is needed.

With PET/CT technology pathophysiological processes at the molecular level can be detected in vivo, which makes it a potentially important research tool to elucidate the mostly unknown mechanisms behind aneurysmal disease. If these processes can be identified and imaged they could serve as a proof of concept when evaluating different interventions to prevent growth and rupture. It has previously been shown in vivo that the chronic inflammation observed in the wall of large asymptomatic AAAs does not have sufficient metabolic activity to be detectable by [18F]-fluorodeoxyglucose (FDG) PET, [11C]PK11195, or [11C]D-deprenyl (50-52). This was also verified in this in vitro study, as neither [11C]DAA1106 nor [11C]DDE showed any substantial binding to AAA tissue.

A limitation of the study is that some of the tested tracers were designed for a totally different specific target, e.g. TF2-IMP. Even though there are no clear indications of CEA receptors in AAA, we found it interesting to test them in an exploratory fashion. The fact that CRPB has such a low affinity (Kd = 5 µM) might be the cause of the negative results. Thus, there might be CRP receptors in the tissue, although at a too low a concentration to be detectable. CRP has previously been found by immunohistochemical means in smooth muscle cells, foam cells, and macrophages in arteriosclerotic aortas (53). Until a CRPB tracer with higher affinity has been developed it is, however, not possible to use CRP as an in vivo target for use in AAA patients.

Macrophages are known to be present in AAA tissue, and as [11C]DAA is a TSPO tracer it might be anticipated that this tracer is a good candidate for visualizing AAA. We have previously shown that the [11C]PK11195 is not capable of detecting the TSPO receptor in vivo in patients with asymptomatic AAAs (52), but even though [11C]DAA has higher affinity than [11C]PK11195, it was not possible to visualize TSPO in AAA with [11C]DAA. To our knowledge no immunohistochemical analysis for detection of TSPO antibody on AAA has been performed. It would be of great interest to investigate a TSPO tracer with even higher affinity than [11C]DAA.

DDE is a tracer with an as yet unknown target receptor, but it has previously been shown to signal for chronic inflammation in vivo (18,19). However, the inflammation in AAA tissue is not active enough for the DDE tracer to become sufficiently taken up.

The DED tracer targets astrocytes, which has been useful in diagnosing Alzheimer’s disease (38) and epilepsy (36,37), while TF2-IMP signals for CEA (40). FMTO visualizes adrenal cortical tumours (41–43), and FVOZ targets the enzyme aromatase (44). The likelihood is low that AAA tissue would have receptors for astrocytes, CEA, mitochondrial cytochrome P-450 species found in the adrenal cortex, or aromatase, but it cannot be entirely ruled out. The fact that these tracers (DED, TF2-IMP, FMTO, and FVOZ) did not signal in AAA tissue makes it unlikely that they are suitable for imaging of the pathophysiological processes of asymptomatic aneurysmal dilatation, at least in their current forms.

There was a substantial specific uptake of [18F]fluciclatide, which targets the integrin αVβ3, an indicator for angiogenesis. Pathological processes in the vasa vasorum and angiogenesis as a cause of AAA formation have previously been suggested (54), and the increase of neovascularization has been demonstrated with immunohistochemical methods (55). With a PET tracer targeting integrin αVβ3 it might be possible to show an increased angiogenesis in vivo. It would then be of interest to study patients with small AAAs and thus to investigate a possible role of changed angiogenetic activities in the formation of AAA.

Conclusion

The integrin-specific tracer [18F]fluciclatide might be useful in the detection of angiogenesis in asymptomatic AAAs. Further in vivo molecular imaging studies of asymptomatic AAAs with [18F]fluciclatide are warranted, and additional in vitro studies of novel PET tracers may reveal new imaging possibilities.