Monitoring the effects of dexamethasone treatment by MRI using in vivo iron oxide nanoparticle-labeled macrophages

Abstract

Introduction: Rheumatoid arthritis (RA) is a chronic disease causing recurring inflammatory joint attacks. These attacks are characterized by macrophage infiltration contributing to joint destruction. Studies have shown that RA treatment efficacy is correlated to synovial macrophage number. The aim of this study was to experimentally validate the use of in vivo superparamagnetic iron oxide nanoparticle (SPION) labeled macrophages to evaluate RA treatment by MRI.

Methods: The evolution of macrophages was monitored with and without dexamethasone (Dexa) treatment in rats. Two doses of 3 and 1 mg/kg Dexa were administered two and five days following induction of antigen induced arthritis. SPIONs (7 mg Fe/rat) were injected intravenously and the knees were imaged in vivo on days 6, 10 and 13. The MR images were scored for three parameters: SPION signal intensity, SPION distribution pattern and synovial oedema. Using 3D semi-automated software, the MR SPION signal was quantified. The efficacy of SPIONs and gadolinium chelate (Gd), an MR contrast agent, in illustrating treatment effects were compared. Those results were confirmed through histological measurements of number and area of macrophages and nanoparticle clusters using CD68 immunostaining and Prussian blue staining respectively.

Results: Results show that the pattern and the intensity of SPION-labeled macrophages on MRI were altered by Dexa treatment. While the Dexa group had a uniform elliptical line surrounding an oedema pocket, the untreated group showed a diffused SPION distribution on day 6 post-induction. Dexa reduced the intensity of SPION signal 50-60% on days 10 and 13 compared to controls (P = 0.00008 and 0.002 respectively). Similar results were found when the signal was measured by the 3D tool. On day 13, the persisting low grade arthritis progression could not be demonstrated by Gd. Analysis of knee samples by Prussian blue and CD68 immunostaining confirmed in vivo SPION uptake by macrophages. Furthermore, CD68 immunostaining revealed that Dexa treatment significantly decreased the area and number of synovial macrophages. Prussian blue quantification corresponded to the macrophage measurements and both were in agreement with the MRI findings.

Conclusions: We have demonstrated the feasibility of MRI tracking of in vivo SPION-labeled macrophages to assess RA treatment effects.

Figure 1

The evolution of superparamagnetic iron oxide nanoparticles (SPION)-labeled macrophages signal on magnetic resonance (MR) images in the presence and absence of dexamethasone (Dexa).(A) T1-weighted images (negative contrast) of SPION-labeled macrophages in the synovium during AIA. SPIONs were administered intravenously on day 5 post-AIA induction. Panel (i-iii) shows representative MR images from a control (untreated) animal with AIA and panel (iv-vi) shows MR images of a Dexa-treated animal on days 6, 10 and 13 post-AIA induction respectively. (B) T1-weighted images of SPION-labeled macrophages post-Gd administration of the same two animals shown in A. SPION signal is seen as a negative contrast, and Gd signal is seen as a positive contrast and depicts synovial edema. Panel (i-iii) shows representative MR images from a control (untreated) animal with AIA and panel (iv-vi) shows MR images of a Dexa-treated animal on days 6, 10 and 13 post-AIA induction respectively. White arrows, SPION signal (dark); red arrows, quadriceps tendon to patella (dark u-shaped line); yellow arrows: femur and tibia; broken yellow arrow, inflamed synovium (bright signal); broken orange line, edema pocket anterior to the femur. Similar results were seen in 13 untreated controls and 15 Dexa-treated animals. AIA, antigen-induced arthritis; Gd, gadolinium chelate.

Figure 2

Qualitative scoring of antigen-induced arthritis (AIA) stage using superparamagnetic iron oxide nanoparticles (SPION)-labeled macrophage and gadolinium chelate (Gd) signal in the synovium on days 6, 10 and 13 post-induction in the presence and absence of dexamethasone (Dexa). (A) T1-weighted MR images showing the diffuseness of the SPION signal in the synovium; MR image with diffuseness (i) score of 3 and (ii) score of 0. (B) Line graph showing the blinded scores of diffuseness of the Dexa-treated and untreated groups. (C) T1-weighted MR images showing the intensity of the SPION signal in the synovium; MR image with intensity (i) score of 3 and (ii) score of 0. (D) Line graph showing the blinded scores of intensity parameter of the Dexa-treated and untreated groups. (E) Post-Gd T1-weighted MR images showing synovial inflammation as a positive contrast; MR image with synovial inflammation (i) score of 3 and (ii) score of 0. (F) Line graph showing the blinded scores of synovial inflammation of the Dexa-treated and untreated groups. All data points are mean ± standard error of mean. B and D: n = 13 untreated controls and n = 15 Dexa-treated group, reduced to 6 and 8 on day 13. B: P=1x10−8 on day 6, D: P = 0.00008 on day 10 and 0.002 on day 13 compared to the untreated control group. F: n = 20 on day 3, n = 14 on day 13, divided between Dexa-treated and untreated. P = 0.00126 on day 6 compared to the untreated controls. Red and yellow arrows indicate the SPION regions of interest where the scoring criteria were assessed. MR, magnetic resonance.

Figure 3

Quantitative measurement of superparamagnetic iron oxide nanoparticles (SPION)-labeled macrophage signal in the presence and absence of dexamethasone (Dexa) using three-dimensional semi-automated segmentation software with difference ultra-short echo time (dUTE) magnetic resonance (MR) images. (A) T1-weighted MR images of SPION-labeled macrophages in the synovium with varying hypointense signals (i) high (ii) low. (B) Corresponding dUTE MR images of the T1-weighted images seen in A with varying hyperintense (positive contrast) signal in the synovium (i) high (ii) low, that were used for three-dimensional signal segmentation. (C) Box plot graph comparing the SPION signal between the untreated and Dexa-treated groups using dUTE images. Data points ± standard error of the mean; n = 23, day 6; n = 22, day 10; n = 14, day 13. P <0.05 compared to the untreated control group (day 6, P = 0.00035; day 10, P = 0.00004; day 13, P = 0.0033).

Figure 4

Photomicrographs of Prussian blue and CD68-immunostained sections confirming in vivo superparamagnetic iron oxide nanoparticles(SPION) labeling of macrophages.(A) Double-stained photomicrographs of (i) Prussian-blue-stained SPIONs (ii) immunofluorescent-stained macrophages using mouse anti-rat CD68 1^ry antibody (as a marker of macrophages) and anti-mouse fluorescein isothiocyanate (FITC) second antibody at 20 times magnification (iii) the overlay of the two previous images. (B, C) Photomicrographs of two separate examples of (i) Prussian blue followed by (ii) CD68-immunostained sections at 40 times magnification. Sections were stained consecutively and micrographs of the same region of interest were obtained.

Figure 5

Photomicrographs of CD68-immunostained sections comparing the evolution of macrophages during antigen-induced arthritis (AIA) in the presence and absence of dexamethasone (Dexa) treatment. Photomicrographs showing macrophages (20 times magnification) on sections stained with a mouse anti-rat CD68 1^ry antibody and a peroxidase goat anti-mouse IgG second antibody. The slides were developed using one step AEC solution. Mayer’s hematoxylin was used as a counter staining. (A, B, C) Photomicrographs from untreated animals (control group) on days 6, 10 and 13 post-AIA induction. (D, E, F) Photomicrographs from Dexa-treated animals on days 6, 10 and 13 post-AIA induction. Red arrows indicate the edema pocket in the knee joint surrounded by inflamed synovial membrane with dense macrophage infiltration.

Figure 6

Quantification of the area and number of CD68-positive macrophages on day 13 post-antigen-induced arthritis (AIA) induction. Photomicrographs of CD68-immunostained sections on day 13 post-AIA induction were scanned and the images were analyzed for the % area of CD68-positive cells (A) and their number (B) using Image J and Tissue Studio® software, respectively. Four sections were quantified and averaged per animal. Data points are mean ± standard error of the mean and n = 4 per group. *P = 0.023 (A) and 0.003 (B) compared to the untreated control group.

Figure 7

Distribution and quantification of Prussian blue stained superparamagnetic iron oxide nanoparticles (SPIONs) on day 13 post-antigen-induced arthritis (AIA) induction. Photomicrographs of Prussian-blue-stained sections showing an example of the distribution of SPIONs (red arrows) in the synovium of untreated animal (A) versus a Dexa-treated animal (B) on day 13 post-AIA induction at 1.5 times magnification. Quantification of the area (C) and number (D) of Prussian-blue-stained SPIONs on day 13 post-AIA induction. Photomicrographs of Prussian-blue-stained sections were scanned and the images were analyzed for the area (C) and the number (D) of SPIONs using Tissue Studio® software. Four sections were quantified and averaged per animal. Data points are mean ± standard error of the mean and n = 5 per group. *P = 0.005 (A) and 0.016 (B) compared to the untreated control group.