¹⁸F-VC701-PET and MRI in the in vivo neuroinflammation assessment of a mouse model of multiple sclerosis

¹⁸F-VC701 was prepared with the automated synthesizer TRACERLAB _FXFN (GE Healthcare, Milan, Italy) as previously described. In brief, [¹⁸F]F⁻ was produced with the ¹⁸O(p,n)¹⁸F nuclear reaction by irradiation of [¹⁸O]water. After an anhydrification step, the labelling was performed with the precursor VC622 (3 mg) in anhydrous DMSO at 140 °C for 20 min. The reaction mixture was injected in semi-preparative HPLC for the purification. The fraction corresponding to ¹⁸F-VC701 was collected in sterile water and the product recovered by solid-phase extraction on pre-activated Sep-Pak tC-18 cartridge. ¹⁸F-VC701 eluted with ethanol (0.7 mL) and diluted with a saline solution (10 mL) for the final formulation. Radiochemical purity of ¹⁸F-VC701 > 98% was requested for mice experiments.

C57BL/6J mice were purchased from Charles River. Age-matched female, 8- to 12-weeks-old, were used in all experimental procedures. Animal were maintained and handled in compliance with the institutional guidelines for the care and use of experimental animals (IACUC), which have been notified to the Italian Ministry of Health and approved by the Ethics Committee of the San Raffaele Scientific Institute (Prot. N. SK552/2012 D.lsg. 116/1992 and N. 722/2016-PR D.lsg. 26/2016).

C57BL/6J mice were immunized subcutaneously with 100 μg MOG_35–55 (MEVGWYRSPFSRVVHLYRNGK, AnaSpec Inc), in incomplete Freud adjuvant (IFA, Sigma) containing 400 μg of heat-inactivated Mycobacterium tuberculosis H37 RA (Difco) and received two intravenous injections (i.v.) of 350 ng of pertussis toxin (List Biological Laboratories Inc), on day 0 and day 2 after immunization. The effect of immunization was checked using a standard clinical scale ranging from 0 to 5 [29] as follows: 0, asymptomatic; 1, limp tail; 2, limp tail and hind limb weakness; 3, partial hind limb paralysis; 4, complete hind limb paralysis; 5, moribund state, death (Fig. 1).

A total of 40 mice were included in this study. For EAE mice: 27 mice were immunized as described in the previous paragraph; one mouse did not manifest clinical signs and was excluded and one had a severe reaction and died after immunization. The remaining 25 EAE mice were included in the study: 15 animals were dedicated to in vivo and ex vivo diagnostic imaging and 10 to immunohistochemistry (IHC) analysis. In particular, seven mice were sacrificed in ex vivo VC701 biodistribution experiment at 14 days post-infection (p.i.); four mice performed the in vivo longitudinal MRI study at 7, 14, and 28 days. The same group of animals underwent a PET scan at day 14 p.i. These animals plus additional 3 performed an ex vivo study with [¹⁸F]VC701 at 30 p.i. Time-activity curve (TAC) study was performed on a separate mouse to define PET acquisition time frame.

For healthy controls, 15 control animals were also included, 9 animals for in vivo and ex vivo imaging and 6 for immunohistochemistry (IHC) analysis.

Disease progression was monitored through longitudinal MRI imaging at different time points after immunization, with T1-Gd and T2-weighted sequences using a small-animal dedicated 7 Tesla MR scanner (Bruker Biospec 30/70, Ettlingen, Germany) and a four-channel mouse-head dedicated phased-array coil.

EAE (n = 4) animals at 7, 14, and 28 days post-immunization and control mice (n = 3) were anesthetized via intra peritoneal (i.p.) injection of tribromoethanol (Sigma Aldrich S.r.l., Italy) at 1.7% (20 μg/g) concentration prior to intravenous (i.v.) administration of Gadobutrol (Gadovist®, Bayer S.p.A., Italy) 0.3 mmol/kg, based on animal weight, in order to assess cerebral lesions. The animals were positioned prone on the dedicated heated apparatus. The image acquisition was performed on coronal plane using the following sequence protocol:

The brain lesions’ volumes were quantified manually using the manufacturer’s software (Paravision 5.1, Bruker, Ettlingen, Germany).

The time course of TSPO signal in EAE mice was evaluated by tissue sampling analysis at 14 (EAE, n = 7) and 30 days post-immunization (EAE, n = 7). EAE animals were compared with 9 aged match healthy animals. Animals were injected in the tail vein with 4.0 ± 0.9 MBq of ¹⁸F-VC701 and sacrificed 240 min later. This time frame was selected on the basis of previous kinetics data on rats [26] and on a brain kinetic study on one EAE mouse showing a slight but progressive increase of brain radioactivity concentration, particularly in the cerebellum that is the brain region more affected in EAE mice. The increase in ¹⁸F-VC701 uptake measured as percentages of the injected dose per gram and cerebellum to cortex ratio is maximum between 2 and 4 h p.i. (see Additional file 1: Figure S1). Plasma was separated by centrifugation and counted with a blood sample in a gamma counter (LKB Compugamma CS1282, Wallac). Immediately after sacrifice, discrete areas as cortex, striatum, hippocampus, cerebellum, cervical enlargement, thoracic, and lumbar spinal cord were sampled and washed with cold saline. Samples were placed in pre-weighed tubes and counted. Radioactivity concentrations were calculated as percentages of the injected dose per gram of tissue (%ID/g) and expressed as ratio between interested area and plasma (%ID/g/%ID/g).

Four EAE mice and 3 healthy controls that underwent the longitudinal MRI study were evaluated also with PET at the time of maximum MRI-based abnormalities, i.e., 14 days post-immunization. ¹⁸F-VC701 was i.v. injected in a tail vein (5.9 ± 0.7 MBq). Each mouse was anesthetized with isoflurane in air (induction, 4%; maintenance, 2%) 240 min later, positioned prone on the PET scanner (YAP-S-PET II, ISE, Italy) bed, and scanned for 30 min (six dynamics frames of 5 min each). For the EAE mouse that underwent kinetic for the definition of the better time frame of tracer study, we acquired brain from ¹⁸F-VC701 i.v. injection up to 240 min. After reconstruction, correction for injected dose, and radioisotope decay, PET images were quantified using dedicated phantom and co-registered with MRI for the analysis with PMOD 3.2v (PMOD Technologies Ltd., Switzerland) software. ¹⁸F-VC701 radioactivity concentration was calculated using region of interest (ROI) analysis. More specifically, circular macro-ROIs were hand-drawn on the forebrain, midbrain, cerebellum, cervical enlargement, thoracic and lumbar spinal cord, and muscle in the MRI images and then applied on the co-registered PET images on three consecutive transaxial slices. Radioactivity concentrations were expressed as a percentage of injected dose per gram (%ID/g) and divided for radioactivity concentration in muscle (%ID/g/%ID/g). The evaluation of the relative distribution of TSPO-PET and MRI (T1 post-gadolinium enhancing and T2 hyperintensities) abnormalities was performed by visual inspection of the co-registered brain images of the four EAE mice.

Mice (5 EAE and 3 control mice per time point) were perfused with PBS followed by 4% paraformaldehyde (PFA). Two-micrometer-thick paraffin-embedded tissue sections from brains and spinal cords have been used for either routine pathological evaluation by hematoxylin and eosin (H&E) staining and immunohistochemistry (IHC). The following primary antibodies were used: rabbit anti-Iba-1 (1:300, Wako), rabbit anti-GFAP (1:200, Dako), and rat anti-MBP (1:50, Chemicon, Thermo Fisher Scientific, USA). Immunostaining has been revealed by Real EnVision Rabbit HRP or Rat-on-Mouse HRP (Biocare Medical, USA) and diaminobenzidine (DAB, Dako, Agilent Technologies, Italy) and counterstained with hematoxylin. Digital images have been acquired by Olympus DP70 camera mounted on Olympus Bx60 microscope, using CellF Imaging software (Soft Imaging System GmbH, Olympus, Italy).

Quantification of neuropathological features, although present both in the brain and spinal cord sections, has been performed in the spinal cords since it is highly reproducible and well established in the literature [30]. The degree of inflammation was expressed as percentage of infiltrated area over the total spinal cord section and demyelination was evaluated by measuring the percentage of demyelinated area over the total white matter area for each spinal cord section. An average of 8–10 spinal cord sections for each mouse was evaluated. Macrophage infiltration/microglial activation have been evaluated by assigning a semi-quantitative score to Iba-1 immunostaining and graded as follows: 0, none; 1, low; 2, moderate; 3, high [30].

Differences in radioactivity concentration between regional distribution of ¹⁸F-VC701 in EAE and control animals were evaluated using the Student t test, defining a statistical significance threshold of p < 0.05. Mann-Whitney test was performed for MRI data of brain lesion volume measured through T1-gadolinium-enhanced and T2-weighted images (Software: GraphPad Prism 5.0).

The vast majority of the immunized mice developed signs of EAE. The disease onset ranged approximately between day 10 and day 11 p.i. with a peak of clinical manifestations at 14–15 days p.i. One mouse did not manifest clinical signs, and one died after immunization. Despite a high variability of response, considering the whole group of mice, disease scores remained fairly stable over time (Fig. 1). At 7 days after disease induction, EAE mice T1 and T2 images showed no differences when compared to the control group, consistent with a previous report [27]. In the acute disease phase (14 days p.i.), lesions could be detected on both T2-weighted and post-contrast T1-weighted sequences in the regions including periventricular white matter, corpus callosum, thalamus, hippocampus, midbrain, pons, and cerebellum (Fig. 2a). At this time point, T2 hyperintense lesions overlapped with Gd-T1 enhancement in the majority of brain regions. Twenty-eight days p.i., signals detected by both T1/T2-weighted images decreased in the majority of regions (Fig. 2b; see panel (c) for control mouse). However, in the periventricular white matter and in the ventral aspect of the hippocampus, T2 alterations were still evident. Overall, these imaging data are consistent with the motor signs of immunized mice. The lesion volume measured on T1-weighted images at 14 days p.i. showed great variability (Fig. 3a), likely mirroring the known variability in the immunization. Belatedly (28 days p.i.), the T1-weighted lesion volume values decreased progressively to zero in all tested mice. The total brain lesion volume on T1-weighted images was lower when compared to that of T2-weighted images (Fig. 3b). T2-weighted images indeed showed the pattern of more concurrent events such as inflammation, edema, and demyelinating process, without however the possibility of differentiating them.

¹⁸F-VC701 allowed detecting ex vivo and in vivo inflammatory reaction in the brain and spinal cord of EAE mice at the time of maximum increase of clinical signs. Mice were analyzed at 4 h after tracer injection to maximize signal to background ratio (see the “Methods” section).

Tissue sampling analysis revealed a significant increase of radioactivity concentration in the EAE mice as compared to control at the level of cortex and cerebellum (p < 0.05), and striatum, hippocampus, cervical enlargement, and thoracic and lumbar spinal cord (p < 0.01) at 14 days p.i. Radioactivity distribution progressed towards a cranio-caudal direction, with lower values in the cortex and higher values in the lumbar trait of the spinal cord (Fig. 4a).

At 30 days p.i., we observed a trend similar to that observed for the acute phase, although significant differences in ¹⁸F-VC701 uptake between EAE and WT (p < 0.05) only in the spinal cord and in discrete brain regions like the cerebellum.

In vivo PET analysis performed 14 days p.i. was consistent with the findings obtained from ex vivo experiments. Macro-ROI analysis showed a progressive increase of ¹⁸F-VC701 uptake towards cranio-caudal direction, with the highest increase of radiotracer uptake localized in lumbar trait of the spinal cord, and significant increase compare to controls observed in midbrain (p < 0.01) and cerebellum, cervical enlargement, and thoracic spinal cord (p < 0.05, Fig. 4c). Coronal and sagittal PET images of the three spinal cord traits considered are shown in Fig. 4d and Additional file 2: Figure S2, respectively. At visual inspection, EAE mice displayed heterogeneous clusters of increased radioactivity uptake in discrete brain regions including anterior cortex, midbrain, cerebellum, cervical enlargement and thoracic spinal cord when compared to control mice (see Figs. 4d and 6 and Additional file 2: Figure S2). Histopathological analysis of spinal cords and brains confirmed the association between the increased clinical score observed in immunized mice and the neuropathological features. Of note, at 2 weeks post-immunization, H&E staining revealed severe acute inflammation, particularly within the spinal cord, and preferentially distributed along the meninges and in perivascular areas (Fig. 5a and left graph). Myelin damage (Fig. 5b) and reactive gliosis (data not shown) were in line with the clinical score. However, during the chronic phase (at 4 weeks post-immunization), inflammation was barely detected (Fig. 5d), whereas demyelination, when quantified on myelin basic protein (MBP) immune-stained sections, was significantly lower as compared to control mice at 2 weeks p.i. (Fig. 5e and middle graph), mainly due to inflammatory clearance and re-myelination. This data mirrored the clinical scores as well. Of note, GFAP immunostaining highlighted diffuse and intense gliosis (data not shown). Accordingly, immunostaining for the microglia/macrophage-specific marker Iba-1 showed that activated microglia and macrophages were highly represented in the spinal cord at the peak of the disease (2 weeks p.i.), while during the chronic phase (4 weeks p.i.), EAE mice showed a significant reduction of the macrophage/microglia infiltration (Fig. 5c, f; right graph). Activated microglia show a typical hypertrophic morphology (inset from panel (c)), while at 4 weeks p.i. microglia shows the classical feature of resting cells with thin ramification and scarce cytoplasm (inset from panel (f)). Iba-1 staining at 14 days revealed the presence of microglial cells/macrophage infiltration in different brain regions including the brain stem, the cerebellum, the hippocampus and the cortex (Fig. 5g). According to the present converging PET, MRI, and ICH data, inflammation and tissue damage were more severe in the spinal cords than in the brain and followed a cranio-caudal gradient, with the lumbar spinal cord sections being the more affected.

PET signals in EAE mice only partially co-localized with damaged brain regions detected with MRI. In one mouse, PET evidence and clinical assessments were negative despite the presence of lesions on T2-weighted images. On the contrary, other mice showed highly concordant PET- and MRI-based evidence in multiple brain regions, i.e., hippocampus (3/4), hypothalamus (2/4), midbrain (3/4), corpus callosum (2/4) pons, and cerebellum both (4/4). However, and particularly for the hippocampus, the distributions of PET and MRI alterations were not completely overlapped (14 days p.i., Fig. 6, Additional file 3: Figure S3).

Otherwise, we observed clusters of significant ¹⁸F-VC701 uptake in absence of MRI T1 or T2 signals in other regions, such as in the striatum, cortex, and cerebellum. Lastly, no increases of radioactivity concentrations were detected in MRI-positive areas such as the thalamus and periventricular white matter (Table 1).