Non-radioactive and sensitive tracking of neutrophils towards inflammation using antibody functionalized magnetic particle imaging tracers

Abstract

White blood cells (WBCs) are a key component of the mammalian immune system and play an essential role in surveillance, defense, and adaptation against foreign pathogens. Apart from their roles in the active combat of infection and the development of adaptive immunity, immune cells are also involved in tumor development and metastasis. Antibody-based therapeutics have been developed to regulate (i.e. selectively activate or inhibit immune function) and harness immune cells to fight malignancy. Alternatively, non-invasive tracking of WBC distribution can diagnose inflammation, infection, fevers of unknown origin (FUOs), and cancer. Magnetic Particle Imaging (MPI) is a non-invasive, non-radioactive, and sensitive medical imaging technique that uses safe superparamagnetic iron oxide nanoparticles (SPIOs) as tracers. MPI has previously been shown to track therapeutic stem cells for over 87 days with a ~200 cell detection limit. In the current work, we utilized antibody-conjugated SPIOs specific to neutrophils for in situ labeling, and non-invasive and radiation-free tracking of these inflammatory cells to sites of infection and inflammation in an in vivo murine model of lipopolysaccharide-induced myositis. MPI showed sensitive detection of inflammation with a contrast-to-noise ratio of ~8-13.

Keywords: antibody; inflammation; magnetic particle imaging; medical imaging; superparamagnetic iron oxide nanoparticles; white blood cells.

Figure 1

Magnetic Particle Imaging using a custom-built vertical bore 6.3 T/m Field- Free Line (FFL) scanner. The MPI scanner allows for 2D projection and 3D tomographic imaging of the spatial distribution of superparamagnetic iron oxide nanoparticles tracers (SPIOs). SPIOs obey Langevin physics; there is high magnetic saturation in response to an applied field and zero coercivity or remanence once the field is removed. In MPI, a time- varying field is applied and only the particles at the FFL flip in response. The flip generates a signal in a receiver coil due to Faraday's law of induction, and rastering the FFL allows the signal to be spatially localized. The FFL scanner can acquire a single 2D projection, or multiple 2D projections at various angles. Classical projection reconstruction algorithms can be implemented to reconstruct 3D MPI images from 2D projections. (2D: two-dimension; 3D: three-dimension; FFL: field-free line; MIP: maximum intensity projection; SPIO: superparamagnetic iron oxide nanoparticles).

Figure 2

Magnetic characterization of the anti-Ly6G SPIOs compared to VivoTrax^TM (VT). Representative transmission electron microscopy (TEM) images (scale bar = 50 nm) were taken of both (A) VT with core diameter = 5.4 ± 1 nm and (B) anti-Ly6G SPIOs with core diameter = 14 ± 2 nm. (C) The magnetization curve of VT and the anti-Ly6G SPIOs as measured in a VSM were fit to a log-normal diameter distribution. Anti-Ly6G SPIOs had a mean magnetic diameter (D_v) of 14.8 nm, with a log-normal standard deviation (σ_{ln d}) of 0.5. VT had (D_v) = 9.8 nm and σ = 0.2. We used a custom arbitrary waveform relaxometer (AWR) to measure the MPI point spread function (PSF) of both samples in saline and in blood. Representative PSFs are shown in (D) for VT and (E) for the anti-Ly6G SPIOs. Note the signal change in the PSF of VT, with the change in signal being significant (p < 0.05). In comparison, anti-Ly6G SPIOs showed stable behavior in both solvents (change in signal was not significant (p = 0.7 > 0.05). (F) A standard curve of MPI signal was taken versus iron concentration for all particles in mouse blood and in saline, with mean and standard error given over n = 3 samples (with 8 acquisition repeats averaged per sample). Both particles show linearity with concentration in both solvents (R² > 0.99). The anti-Ly6G SPIOs showed significantly better resolution (p < 0.01) and sensitivity (1.8 ± 0.3 times better, with p < 0.01) than VivoTrax^TM. VivoTrax^TM showed a significant (p < 0.01) change in sensitivity (V/g) in different media, while the anti-Ly6G SPIOs did not (p > 0.05). This standard curve was also repeated in the scanner and is included in Figure 3. (H: magnetic field; M: Magnetization; MPI; magnetic particle imaging; PSF: point-spread function; SPIO: superparamagnetic iron oxide; VT: VivoTrax^TM).

Figure 3

Standard Curve of the anti-Ly6G SPIOs and VivoTrax^TM (VT) in our Field-Free Line Scanner. (A) the images of the standard curve for anti-Ly6G SPIOs and VT in saline are shown in linear heat scale scaled to the maximum signal between both curves (FOV = 4.1 x 6.2 cm², scan time = 37s). Note the increased conspicuity of anti-Ly6G SPIOs when compared to VivoTrax^TM (B) The standard curves in saline (given as the sample SNR averaged from the 5 pixels by 5 pixels center of each sample), alongside standard curves in blood. The linear fits to each particle and solvent pair are mostly good (R² > 0.99), except for VT in blood (R² = 0.85) - this can be attributed to experimental error in the sample at 0.7 μg mm^-2. Note the similarity to Figure 2F, with the sensitivity of the anti-Ly6G SPIOs being 1.7 ± 0.5 times better than that of VivoTrax^TM. (SNR: signal-to-noise ratio; SPIO: superparamagnetic iron oxide; VT: VivoTrax^TM).

Figure 4

Labeling of immune cells using anti-Ly6G SPIOs. (A) TEM images of whole blood sample incubated with anti-Ly6G SPIOs (scale bar = 2 μm). (B), (C) Magnified TEM images (scale bar = 200 nm) & (D) (scale bar = 100 nm) of yellow box marked in (A). The arrows point to the blood cell-membrane bound nanoparticles. The membrane bound particles had a core diameter of 16 ± 5 nm, corresponding well with the size of anti-Ly6G SPIO particles observed in Figure 2 (SPIO: superparamagnetic iron oxide (nanoparticles); TEM: transmission electron microscopy).

Figure 5

Cell enrichment analysis of MPI tracers using flow cytometry: Flow cytometry data represented as % enriched immune cell population (normalized to total cells counted) in the “Eluate” and “Column” content collected during and after magnetic separation of RBC lysed mouse blood using anti-Ly6G SPIOs, anti-F4/80 SPIOs and VivoTrax^TM. The control data represents the % of immune cells of interest (normalized to total cells counted) in unseparated RBC lysed blood samples. We were particularly interested in the granulocyte population (of which neutrophils are the largest subset) and the CD11b⁺ & Ly6G⁺ population. A significant enrichment of both granulocyte population and CD11b⁺ & Ly6G⁺ was observed in the “Column” using anti-Ly6G SPIOs. Anti-F4/80 SPIOs showed no enrichment of granulocytes or neutrophil population, which is expected due to lack of specificity to those populations. Finally, Vivotrax^TM showed significant enrichment of both granulocytes and neutrophil population (P<0.05 (**) and P<0.001 (***)) but the percentage relative enrichment in the “column” was not significant with respect to control (P> 0.05) (MPI: magnetic particle imaging; RBC: red blood cells; SPIO: superparamagnetic iron oxide).

Figure 6

Immuno-MPI: Anterior-posterior maximum intensity projection (MIP) images of 3D MPI data, 24 hours post anti-Ly6G antibody SPIO tracer administration in (A) a healthy mouse (B) mouse with LPS endotoxin induced myositis in the right leg (MIP: maximum intensity projection; MPI: magnetic particle imaging; SPIO: superparamagnetic iron oxide).

Figure 7

Inflammation images in three different mouse subjects with myositis and i.v. administered anti-Ly6G-SPIO. Mouse 1 had a cardio-pulmonary reaction to the tracer and the experiment had to be terminated early (1-hour post tracer administration), while Mouse 2 and Mouse 3 were acquired at 24 hours post tracer administration. The CNR of the site of inflammation was between 8-13 (CNR: contrast-to-noise ratio; SPIO: superparamagnetic iron oxide).

Figure 8

Inflammation images in three different mouse subjects with myositis and using VivoTrax^TM as tracer. At a dose of 5 mg of Fe/kg the CNR was ~1-2 at the site of myositis acquired 24 hours post-tracer administration (CNR: contrast-to-noise ratio).

Figure 9

Change in circulation times of tracer after inducing inflammation. In healthy mice (top row), anti-Ly6G SPIO tracer predominantly distributes in the organs of the RES including the liver, spleen, and bone marrow, at around 4-5 hours post-tracer administration. We observed a change in the circulation times of the tracer in myositis induced mice (bottom row). These mice showed prominent MPI signal from the ventricles of the heart around 4-5 hours of post-tracer administration indicative of longer circulation times of tracer post inflammation (RES: reticuloendothelial system; SPIO: superparamagnetic iron oxide).

Figure 10

Confirmation of inflammation. (Left) The neutrophil activity at the site of myositis was confirmed by using i.p. administered luminol, which provides bioluminescent visualization of myeloperoxidase (MPO) enzyme activity. (Right) Histological analysis of anti-Ly6G nanoparticle uptake in mouse models. Images were captured by brightfield microscopy following staining of infected tissues using H&E staining for tissue morphology, Prussian blue staining for SPIO, and immunohistochemistry for myeloperoxidase. The Prussian blue confirmed the distribution of anti-Ly6G SPIO in bone marrow in regions with increased MPO enzyme expression (myeloid cells). The images shown in the bottom row are muscle tissues from the right infected leg. Inflammation injury was observed in the H&E stain (arrows), with the corresponding regions showing areas of increased iron and MPO enzyme activity (H&E: hematoxylin & eosin; MPO: myeloperoxidase; SPIO: superparamagnetic iron oxide (nanoparticles).

Figure 11

Cell Enrichment Analysis of MPI tracers using Flow Cytometry (A) Cells obtained after RBC lysis of mouse blood were incubated with SPIO of interest. The incubated cells were sent through a magnetic column, and unlabeled cells were collected as “eluate” from the column. After removing the magnet utilized for separation, the column was further flushed, and the contents were collected as the “column” subset of cells. The immune cell population from the “eluate” was evaluated for degree of depletion by the SPIO relative to the extent to which it is enriched from the “column” in comparison with control (untouched RBC lysed blood) using flow cytometry. The following gating procedure was carried out on the flow cytometry data: (B) the propidium iodide channel was used to gate live cells, which are negative for the dye, (C) the granulocyte population of immune cells were chosen utilizing the side scatter and forward scatter profile and (D) the cells were analyzed for neutrophil cell marker expression (anti-Ly6G-PE and CD11b-FITC) utilizing the PE and FITC channels. Untouched (i.e. unseparated) samples of RBC lysed blood were used as a control to gauge the native proportion of blood cells in the samples (MPI: magnetic particle imaging; RBC: red blood cells; SPIO: superparamagnetic iron oxide).