Multimodal Imaging with NanoGd Reveals Spatiotemporal Features of Neuroinflammation after Experimental Stroke

Abstract

The purpose of this study is to propose and validate a preclinical in vivo magnetic resonance imaging (MRI) tool to monitor neuroinflammation following ischemic stroke, based on injection of a novel multimodal nanoprobe, NanoGd, specifically designed for internalization by phagocytic cells. First, it is verified that NanoGd is efficiently internalized by microglia in vitro. In vivo MRI coupled with intravenous injection of NanoGd in a permanent middle cerebral artery occlusion mouse model results in hypointense signals in the ischemic lesion. In these mice, longitudinal two-photon intravital microscopy shows NanoGd internalization by activated CX3CR1-GFP/+ cells. Ex vivo analysis, including phase contrast imaging with synchrotron X-ray, histochemistry, and transmission electron microscopy corroborate NanoGd accumulation within the ischemic lesion and uptake by immune phagocytic cells. Taken together, these results confirm the potential of NanoGd-enhanced MRI as an imaging biomarker of neuroinflammation at the subacute stage of ischemic stroke. As far as it is known, this work is the first to decipher the working mechanism of MR signals induced by a nanoparticle passively targeted at phagocytic cells by performing intravital microscopy back-to-back with MRI. Furthermore, using a gadolinium-based rather than an iron-based contrast agent raises future perspectives for the development of molecular imaging with emerging computed tomography technologies.

Keywords: intravital two-photon microscopy; magnetic resonance imaging; microglia/macrophage; multimodal nanoprobe; neuroinflammation; stroke.

Figure 1

NanoGd is internalized by microglial cells in vitro without increasing proinflammatory cytokine production. A) Confocal images of Iba‐1‐stained microglial cells incubated A1) without or A2,A3) with NanoGd (0.5 mmol L^–1; 1.5 mmol L^–1). White arrowheads indicate Iba1+ cells that colocalized with NanoGd (scale bars: 50 µm for overview images; 10 µm for magnified insets). B) Percentage of cultured Iba1+ cells that internalized NanoGd. Data are displayed as mean ± SD. Independent experiments, n = 2; replicates per condition, n = 3 for each experiment. Significant differences between experimental conditions were calculated on one‐way ANOVA and are indicated by *** for p < 0.001. C) Quantification of interleukin‐6 (IL‐6) and tumor‐necrosis factor alpha (TNF‐α) production by microglial cells exposed to different concentrations of NanoGd. Cells exposed to LPS but not to NanoGd (condition “0 × 10⁻³ m NanoGd + LPS”) were used as a positive control for production of proinflammatory cytokines. Independent experiments, n = 2; replicates per condition, n = 3 for experiment 1 and n = 2 for experiment 2. On boxplots, the “X” represents the mean and the circles show the interior or outlier points.

Figure 2

Experimental timeline and study design. A) Experimental timeline designs for NanoGd in vivo multimodal imaging. 2γ 1: intravital two‐photon microscopy, imaging session 1; 2γ 2: intravital two‐photon microscopy, imaging session 2. B) Summary of imaging exams undergone by each mouse, in 3 groups. MRI 1 represents the baseline pre‐NanoGd session, and MRI 2 the post‐NanoGd MRI session. Time intervals in hours (24, 30, 54, or 72 h) correspond to the exact time of imaging post‐pMCAo. Asterisks correspond to changes in animal number of group I and group III: ^* n = 6 and ^** n = 1, both due to the death of one animal during the first two‐photon microscopy session. Of note, statistical analyses have been performed for experimental groups with a sample size superior to n = 2. i.v.: intravenous; w/o: without.

Figure 3

In vivo multiparametric MRI allows the characterization of NanoGd distribution in the ischemic lesion. A,B) Pre‐ and C,D) post‐NanoGd MRI for three representative mice, one pMCAo and one sham‐operated mouse, both injected with NanoGd, and one noninjected pMCAo mouse. For each MRI sequence, only one transversal slice is shown (scale bars: 1 mm). A) T2‐weighted images (T2‐WI) show an ischemic lesion in pMCAo mice (dotted white lines), but not in the sham mouse. Black arrowheads indicate hypointense artefacts associated with pMCAo surgery. B) BBB disruption was assessed on T1‐weighted images (T1‐WI), pre‐ and postinjection of a small gadolinium chelate. T1 enhancement in pMCAo mouse brain was indicative of BBB disruption (white arrowheads). 48 h following NanoGd injection, NanoGd presence at the ischemic lesion was observed on C) T2‐WI and D) T2*‐WI. Red arrowheads indicate hypointense signals in the ischemic lesion of the pMCAo mouse. E) Percentage of T2 hypointense signals inside the ischemic lesion of mice subjected to pMCAo but not injected with NanoGd (group II, n = 4 mice) on baseline and post‐NanoGd (D3) MRI. F) Percentage of T2 hypointense signals inside the ischemic lesion of mice subjected to pMCAo and injected with NanoGd (group I, n = 14 mice). On these boxplots, the “X” represents the mean and the circles show the interior or outlier points. Significant differences between experimental groups on paired Student's t‐test are indicated by *** for p < 0.001. Gd‐DOTA: Dotarem. w/o: without. Ns: nonsignificant.

Figure 4

Visualization of NanoGd interaction with CX3CR1‐GFP/+ cells in the ischemic brain with intravital two‐photon microscopy. A) Schematic representation of the ischemic brain for a pMCAo mouse. The three areas imaged with two‐photon microscopy are represented (green boxes): 1 = extralesional area; 2 = border zone; 3 = ischemic core. B–D) Representative images of two‐photon microscopy for a pMCAo mouse injected with NanoGd, for the first (D1) and the second imaging session (D2). For each subregion, images on the top present fluorescence signals from CX3CR1‐GFP/+ cells (in green) and NanoGd (in red), images in the middle represent fluorescence signals from NanoGd only (in red), and images on the bottom show colocalization between CX3CR1‐GFP/+ cells and NanoGd signal (in yellow). Examples of highly ramified CX3CR1‐GFP/+ cells are indicated by light blue arrowheads, and white arrowheads indicate less‐ramified/large body CX3CR1‐GFP/+ cells. (Scale bar: 20 µm.)

Figure 5

Two‐photon quantitative analyses reveal NanoGd brain diffusion and internalization by CX3CR1‐GFP/+ cells in the ischemic lesion. Quantification of NanoGd brain diffusion and cell internalization at D1 and D2 for pMCAo mice with complete two‐photon intravital imaging: extralesional area, border zone, and ischemic core (group I, n = 4). A) Graphic representation of NanoGd fluorescence intensity inside brain interstitium. The “X” represents the mean and the circles show the interior or outlier points. Number of B) parenchymal and C) circulating CX3CR1‐GFP/+ cells that internalized NanoGd (CX3CR1‐GFP+/NanoGd+ cells, light gray) or not (CX3CR1‐GFP+/NanoGd− cells, dark gray). Data are displayed as mean ± SD and the percentage of CX3CR1‐GFP/+ cells that internalized NanoGd is also indicated. For all graphs, significant differences between brain areas and imaging days were calculated on two‐way ANOVA and are indicated as *** for p < 0.001. There was no significant effect of imaging day (D1 vs D2). For the “NanoGd uptake” graphs, statistical analysis was performed on the CX3CR1‐GFP+/NanoGd+ cell percentage. a.u.: arbitrary unit.

Figure 6

Spatial distribution of NanoGd following pMCAo is confirmed by ex vivo analysis. A) Hypointense signals on in vivo T2‐WI colocalized with hyperintense signals observed on B) maximum intensity projection (MIP) obtained from ex vivo X‐ray phase‐contrast image, used as gold standard to map Gd distribution. Red arrowheads indicate these signals in the ischemic lesion. Both images were obtained from the same pMCAo mouse injected with NanoGd (scale bar: 1 mm). C,D) Fluorescence microscopy images obtained from brain sections of a representative pMCAo mouse. C) The ischemic lesion is delineated by a dotted white line on a macroscopic view of the ischemic hemisphere (scale bar: 250 µm). D) Higher magnification shows CX3CR1‐GFP/+ cells and NanoGd in the D1) extralesional area, D2) border zone, and D3) ischemic core. White arrowheads indicate the area of colocalization between CX3CR1‐GFP/+ cells and NanoGd (scale bars: 50 µm for overview images; 20 µm for magnified insets).

Figure 7

Characterization of NanoGd subcellular distribution using transmission electron microscopy. A–C) Transmission electron microscopy (TEM) images from three brain areas of a pMCAo mouse injected with NanoGd: extralesional area, border zone, and ischemic core. A) Focus on an immune phagocytic cell (dark cell) in the extralesional area. Higher magnification shows absence of electron‐dense structures inside the lysosomes (scale bar: 2 µm for overview image; 200 nm for magnified inset) B) Isolated electron‐dense structures detected within the interstitial tissue of the border zone, indicated by red arrows (scale bar: 100 nm). Images from the ischemic core show accumulation of single and aggregated electron‐dense structures C) inside the interstitial space (red arrows; scale bar: 100 nm), D) inside phagocyte lysosomes (scale bar: 0.5 µm for overview image; 100 nm for magnified inset), and E) inside the endothelial cells (EC) that surround cerebral vessels (scale bar: 1 µm). Higher magnification of the cerebral vessel wall (magnified insets in (E)) shows endothelial cells and their basal lamina membrane (black arrowheads), which delineates the extremity of the cerebral endothelium (scale bar: 200 nm). In these magnified insets, NanoGd is detected in endothelial cell cytoplasm (red arrowheads) inside cytoplasmic vesicles identified by the presence of biological membrane (black arrows), and outside endothelial cells as indicated by red arrowheads (panels (B) and (C)). F) Higher magnification of the ischemic core TEM image shows that the maximal height of the electron‐dense structure is around 25 nm.