Near Infrared Fluorescence (NIRF) Molecular Imaging of Oxidized LDL with an Autoantibody in Experimental Atherosclerosis

We aimed to develop a quantitative antibody-based near infrared fluorescence (NIRF) approach for the imaging of oxidized LDL in atherosclerosis. LO1, a well- characterized monoclonal autoantibody that reacts with malondialdehyde-conjugated LDL, was labeled with a NIRF dye to yield LO1-750. LO1-750 specifically identified necrotic core in ex vivo human coronary lesions. Injection of LO1-750 into high fat (HF) fed atherosclerotic Ldlr−/− mice led to specific focal localization within the aortic arch and its branches, as detected by fluorescence molecular tomography (FMT) combined with micro-computed tomography (CT). Ex vivo confocal microscopy confirmed LO1-750 subendothelial localization of LO1-750 at sites of atherosclerosis, in the vicinity of macrophages. When compared with a NIRF reporter of MMP activity (MMPSense-645-FAST), both probes produced statistically significant increases in NIRF signal in the Ldlr−/− model in relation to duration of HF diet. Upon withdrawing the HF diet, the reduction in oxLDL accumulation, as demonstrated with LO1-750, was less marked than the effect seen on MMP activity. In the rabbit, in vivo injected LO1-750 localization was successfully imaged ex vivo in aortic lesions with a customised intra-arterial NIRF detection catheter. A partially humanized chimeric LO1-Fab-Cys localized similarly to the parent antibody in murine atheroma showing promise for future translation.


Supplementary Figure 2. In vivo aortic imaging of LO1-750 in Ldlr
-/mouse model using hybrid IVIS Spectrum/ CT. (A) LO1-750 localizing to the ascending aorta and aortic arch when scanning the thoracic ROI (red) of a one year old Ldlr -/mouse fed a HF diet for 42 weeks; (B) litter mate control showing very little uptake in the same areas following iv injection of equivalently-labeled isotype control antibody (IgG3-750); (C,D) demonstrates another example, with LO1-750 adjacent to calcium seen on CT at the aortic arch (red arrow) seen on CT (C), co-localizing with LO1-750 signal (green arrow) in (D). LO1-750 also identified another hotspot (blue arrow), which was demonstrated to be in the left subclavian artery upon cross-sectional imaging. (E) Macroscopic ex vivo NIRF of the aorta with its surrounding fatty tissue demonstrates the LO1-750 signal is absent in the fat surrounding the aorta with hot spots in the aortic arch (white arrow), and abdominal aorta (green arrows).

Supplementary Movie 3. Subendothelial localization of LO1-750 following iv injection.
An Ldlr-/mouse was injected iv with LO1-750 and phycoerythrin -conjugated anti-CD31, and 4 hr later the aorta was removed for ex vivo analysis by confocal microscopy. The movie shows that LO1-750 (red) localized beneath endothelium, as detected by anti-CD31 (green).

LO1 and control IgG3
IgG3 control antibody (I5654, Sigma-Aldrich, UK) was shown by ELISA not to react with LDL or MDA-LDL. Antibodies were buffer-exchanged into PBS with Zeba Desalt Spin Columns (Pierce, Rockfor, IL), and then concentrated to 2 mg/ml prior to labeling, using 100,000 NMWL Centricon centrifugal filter devices (Millipore, Watford, UK). Integrity of the antibodies was monitored using SGS-PAGE followed by Colloidal Blue (Invitrogen, Paisley, UK) staining. An ELISA was used to confirm functional reactivity of LO1 before and after labeling, using MDA-LDL or anti-LO1 idiotype H3 on the solid phase, as previously 1 .

Details of the chimeric LO1-Fab-Cys construct
We molecularly expressed the chimeric Fab construct of LO1 with a cysteine-tagged heavy chain (LO1-Fab-Cys). Constructs consisting of human CH1 and CL regions fused respectively to LO1 VH and VL DNA sequences were each cloned into a mammalian expression vector backbone in frame with a DNA sequence encoding an N-terminal secretory signal peptide. The vectors were transfected into HEK293/6E cells and maintained in suspension to express secreted LO1-Fab-Cys molecules. Supernatants were harvested and clarified after 7 days by centrifugation and passing through a 0.22 µm filter. The samples were then concentrated 10-fold using tangential flow prior to purification. Purification was undertaken using an anti-human CH1-IgG purification column (BAC AV, Germany). The purity of LO1-Fab-Cys was confirmed with gel electrophoresis and colloidal blue staining under reducing and non-reducing conditions. Mass spectrometry was used to confirm the mass of the purified LO1-Fab-Cys. Retention of antigen-binding function was confirmed by an ELISA testing the ability of immobilized LO1-Fab-Cys to capture biotinylated anti-LO1 idiotype H3 1 . Inactivation of LO1-Fab-Cys to obtain a negative control was achieved by blocking critical amines required for antigen binding. To achieve this, LO1-Fab-Cys was labeled with an amine reactive dye (Vivo Tag750) to an equivalent degree of labeling (DOL) as with the malemide reactive dye. Deactivation was tested by ELISA as before.

Confocal microscopy on mouse tissues
Tissues were stored at -80°C before use and cryosectioned as previously described 2 . En face preparations of aortae were prepared by harvesting the aortae, and cutting them longitudinally. They were then thoroughly cleaned, fixed in 2% formalin, premeabilized by incubating in 0.5% triton to five min, washed and mounted en face with the lumen facing up.
When staining with antibodies labeled with VivoTag-S 750-MAL, tissues were blocked with 2% bovine serum albumin (BSA), washed and incubated with antibodies (10µg/ml) for 1 hr at 4°C in a humidified chamber. This was followed by a wash using PBS and finalized by nuclear staining with SYTO-24. The slides were then washed and glass cover slips mounted using Fluoromount G. We optimized a Leica SP5 MP inverted confocal microscope for NIRF imaging at 750 nm. The 633nm laser power was set to 93% or above, and a narrow range (750-800nm) was selected for the emission level to eliminate autofluorescence signal. Most images were obtained using a 10x microscope lens with the frame average of 4.

Confocal microscopy on human tissue
Human carotid endarterectomy specimens were freshly collected with consent and Furthermore, well-characterized advanced and intermediate lesions from human coronary sections were also studied. As specified in the AtheroExpress protocol 3 , histological sections were classified according to overall appearance into: "atheromatous lesions" containing a large lipid core (>40% of plaque area), high macrophage infiltration with low smooth muscle cell and collagen content, "fibrous lesions" with a small (<40%) or absent lipid core, low macrophage content and high smooth muscle cell and collagen content, and "fibrousatheromatous lesions" as an intermediate between the two other phenotypes.

Ex vivo confocal analysis of antibody targeting in mice
Following intravenous injection of labeled antibodies, animals were euthanized and the aortae collected and processed as above. The relation of injected antibody to macrophages was obtained by treating aortae with 0.5% triton for 1 hr, followed by staining with rat anti-mouse Macrophages/Monocytes (MOMA-2):Alexa Fluor® 647 (Serotec, Oxford, UK). In some animals, localization of injected LO1 or control IgG3 in relation to endothelium was studied by injecting phycoerythrin (PE)-conjugated anti-CD31 antibody (Biolegend, San Diego, CA) 10 min prior to humanely killing. To minimize interference between different fluorochromes, each channel was acquired in sequence. To obtain 3D images of full aortae or ROIs, images were acquired as above, with Z stacking and tile scanning. Image analysis for all IHC studies was undertaken using Volocity® 3D Image Analysis Software (PerkinElmer, Massachusetts USA).

IHC in the rabbit model
The aortic arch was removed and fixed in 4% phosphate buffered formaldehyde and then wax-embedded. Briefly, serial 3 μm paraffin sections were dewaxed and rehydrated.
Endogenous peroxidase activity was inhibited by incubation with 3% (v/v) hydrogen peroxide. After blocking sections with 20% (v/v) goat serum in PBS, sections were incubated overnight at 4 o C with either 20µg/ml LO1, mouse monoclonal antibody against α-smooth muscle actin (Sigma, UK), or mouse monoclonal antibody against rabbit macrophages (RAM11) (Dako, UK), diluted in 1% (w/v) BSA in PBS. Cell nuclei were visualized with haematoxylin. A negative control, where the primary antibody was replaced with mouse IgG at the same dilution, was always included.

Ex vivo fluorescence microscopy on rabbit aortae
Fluorescence microscopy of plaque and normal vessel sections was performed on adjacent sections from fresh frozen rabbit aortae as previously described) 45 . Using an upright epifluorescence microscope (Nikon Eclipse 90i; Tokyo, Japan), fluorescence images were obtained in the NIR channel for FTP11-Cy7 (excitation/emission 710/810 nm; exposure time 50 ms), and FITC channel for autofluorescence (excitation/emission 480/535 nm; exposure time 50 ms 6 .

IVIS Spectrum fluorescence imaging and CT (IVIS/CT) in mice
Mice were anesthetized with isoflurane using a vaporizer, and 3D fluorescence images were acquired through Fluorescence Imaging Tomography (FLIT; IVIS® Spectrum, Caliper LifeSciences). We used a GFP filter set (excitation wave length, 710nm ± 15 nm; emission wave length, 820 nm± 15 nm) to detect antibodies conjugated with VivoTag-S 750-MAL.
Mice were then transferred anaesthetized in the imaging cassette to the micro-CT imaging suite (Inveon PET-CT, Siemens). CT images were obtained (100µm, 80KVp, 500 µAmp, 120 projections, exposure time 220 ms). Contrast-enhanced high resolution CT localized the aortic root and major arteries in the neck and thorax to guide interpretation. For blood distribution studies, serial tail bleeds at intervals up to 68 hours were performed and samples mixed with equal volumes of heparinized PBS. Blood samples were imaged on a black plate for epifluorescence (ex710, em820) and measured units expressed as total radiant efficiency.
Quantification was obtained by non-linear regression curve fit to a known dilution series of the agent. For organ distribution studies, tissues were harvested and imaged in an opaque plate as above. SNR was calculated and expressed as mean + SEM.

Fluorescence molecular tomography-computed tomography (FMT/CT) in mice
FMT was acquired with an FMT 4000 fluorescence tomography imaging system (PerkinElmer, Massachusetts USA), which is equipped with four lasers and allows acquisition of NIRF from four tracers with distinct excitation and emission wavelengths.

Two-dimensional NIRF imaging device and catheter for the rabbit model
The two-dimensional NIRF imaging device and catheter apparatus has been fully described previously 5 . Briefly, the intravascular optical probe is capable of performing an over-the-wire pull back intravascular acquisition using a 750 nm laser light and collects the subsequent NIRF emission. The fiber is rotated and translated using mechanical stages to collect fluorescence and generate a 2D NIRF image with longitudinal and angular coordinates.

Intravascular NIRF and IVUS imaging in rabbits
Animals were anesthetized as described previously 5 . A 5-F introducer was inserted into the right carotid artery using fluoroscopic and angiographic guidance. Iodinated contrast was injected and baseline x-ray angiography was recorded using standard cineangiography. An