Enhanced cerebrovascular permeability and cellular infiltration mark the onset of early multiple sclerosis lesions. So far, the precise sequence of these events and their role in lesion formation and disease progression remain unknown. Here we provide quantitative evidence that blood–brain barrier leakage is an early event and precedes massive cellular infiltration in the development of acute experimental allergic encephalomyelitis (EAE), the animal correlate of multiple sclerosis. Cerebrovascular leakage and monocytes infiltrates were separately monitored by quantitative in vivo MRI during the course of the disease. Magnetic resonance enhancement of the contrast agent gadolinium diethylenetriaminepentaacetate (Gd‐DTPA), reflecting vascular leakage, occurred concomitantly with the onset of neurological signs and was already at a maximal level at this stage of the disease. Immunohistochemical analysis also confirmed the presence of the serum‐derived proteins such as fibrinogen around the brain vessels early in the disease, whereas no cellular infiltrates could be detected. MRI further demonstrated that Gd‐DTPA leakage clearly preceded monocyte infiltration as imaged by the contrast agent based on ultra small particles of iron oxide (USPIO), which was maximal only during full‐blown EAE. Ultrastructural and immunohistochemical investigation revealed that USPIOs were present in newly infiltrated macrophages within the inflammatory lesions. To validate the use of USPIOs as a non‐invasive tool to evaluate therapeutic strategies, EAE animals were treated with the immunomodulator 3‐hydroxy‐3‐methylglutaryl Coenzyme A reductase inhibitor, lovastatin, which ameliorated clinical scores. MRI showed that the USPIO load in the brain was significantly diminished in lovastatin‐treated animals. Data indicate that cerebrovascular leakage and monocytic trafficking into the brain are two distinct processes in the development of inflammatory lesions during multiple sclerosis, which can be monitored on‐line with MRI using USPIOs and Gd‐DTPA as contrast agents. These studies also implicate that USPIOs are a valuable tool to visualize monocyte infiltration in vivo and quantitatively assess the efficacy of new therapeutics like lovastatin.
Multiple sclerosis is a chronic inflammatory demyelinating disease of the CNS, characterized by the presence of sclerotic lesions or plaques scattered throughout the brain (Ewing and Bernard, 1998). An early event in the development of inflammatory lesions during multiple sclerosis is the formation of cellular infiltrates, consisting mainly of monocyte‐derived macrophages (De Vries et al., 1997; Lassmann et al., 1997; Al Omaishi et al., 1999). In particular, these infiltrated monocytes are central to the process of myelin degradation and tissue damage that characterizes multiple sclerosis (Brück et al., 1996). Reports in either multiple sclerosis patients or in animal models for multiple sclerosis suggest that increased permeability of the blood–brain barrier (BBB) is also an early phenomenon in plaque development and may even initiate lesion formation, although clear evidence is lacking (Kermode et al. 1990; Koh et al. 1993; Moor et al., 1994; De Vries et al., 1997; Werring et al. 2000).
For proper monitoring of disease progression, new non‐invasive methods are required since traditional immunohistochemical techniques can only be applied after fixation of the tissue. Conventional MRI has become established as the imaging modality for diagnosing multiple sclerosis by visualization of neuropathological changes in multiple sclerosis patients (McDonald et al., 2001) and is recommended as a primary outcome measurement in early phase clinical trials (Miller et al. 1996; Filippi et al. 2002; Sormani et al. 2002). Gadolinium diethylenetriaminepenta‐acetate (Gd‐DTPA) enhanced MRI is a powerful diagnostic tool to detect a defective function of the BBB and new lesion formation in multiple sclerosis patients (Brück et al., 1997). Gd‐DTPA enhancement, however, provides no direct information about the trafficking of mononuclear cells into the brain parenchyma (Kermode et al., 1990; Katz et al., 1993; Tofts, 1997; Kappos et al., 1999; Silver et al., 1999). Recently, contrast agents based on ultra small particles of iron oxide (USPIOs) have been developed for MRI. These particles accumulate in cells with phagocytic function (Weissleder et al., 1990; Seneterre et al., 1991), and it has been shown that USPIOs are detectable by MRI at sites of cellular activity during neuroinflammatory processes like experimental allergic encephalomyelitis (EAE), the animal model for multiple sclerosis (Xu et al., 1998; Dousset et al., 1999a, 1999b, 2002), and ischaemia (Rausch et al., 2001, 2002). Therefore, it is suggested that the presence of USPIOs in the brain may reflect infiltration of phagocytes filled with USPIOs, although there are so far no data that support this assumption. The major goal of this study is therefore to elucidate the temporal events of BBB impairment and mononuclear cell infiltration in an acute model of EAE using Gd‐DTPA and USPIO‐enhanced MRI.
Various therapies for multiple sclerosis have focused on the inhibition of cellular infiltration into the brain parenchyma either by direct blocking adhesion molecule interaction (Yednock et al., 1992; Baron et al., 1993; Huitinga et al., 1993; Gordon et al., 1995; Engelhardt et al., 1998; Brocke et al., 1999; Miller et al., 2003) or by diminishing cellular diapedesis by immunomodulation as has been shown for interferon‐β (Corsini et al., 1997; Lou et al., 1999; Kallmann et al., 2000; Floris et al., 2002). Recently, the family of the 3‐hydroxy‐3‐methylglutaryl Coenzyme A reductase inhibitors, the so‐called statins, has been put forward as an attractive candidate for multiple sclerosis therapy (Stanislaus et al., 1999, 2001, 2002; Kwak et al., 2000; Neuhaus et al., 2002; Youssef et al., 2002). Treatment of animals with chronic or acute EAE with statins ameliorated clinical signs and inhibited cellular infiltration, especially of monocytes, into the CNS (Stanislaus, et al., 1999, 2001; Youssef et al., 2002). To validate the use of USPIO enhanced imaging as a tool to determine the effect of therapeutic agents that are designed to halt infiltration of monocytes into the CNS, we treated a subgroup of animals with lovastatin.
Here we show for the first time that the contrast agents Gd‐DTPA and USPIOs can be used to quantitatively visualize two distinct neuropathological processes in the course of the disease. Our data indicate that BBB leakage precedes monocyte infiltration into the CNS, implicating a potential sequence of events in neuropathological lesion formation. Moreover, USPIOs promise to be a valuable tool to monitor cellular infiltration in vivo and to evaluate therapeutic efficacies of drugs on this process.
Material and methods
Induction of acute EAE
Acute EAE was induced in 8–11 week old male Lewis rats (190–210 g; Harlan CPB, Zeist, The Netherlands), which were kept under standard laboratory conditions. All experimental procedures were approved by the Ethical Committee for Animal Experiments of the VU Medical Centre, Amsterdam, The Netherlands.
Acute EAE was induced according to Floris et al. (2002). At day 0, rats were injected subcutaneously in one hind footpad with 20 µg guinea pig myelin basic protein (MBP) in phosphate‐buffered saline (PBS) mixed with complete Freund’s adjuvant (CFA; 4 mg/ml mycobacterium tuberculosis H37Ra; Difco Laboratories, Detroit, MI, USA). Control animals were injected with PBS mixed with CFA. Neurological aberrations were scored daily and graded from 1 to 5: 0 = no neurological abnormalities; 0.5 = partial loss of tail tonus; 1 = complete loss of tail tonus; 2 = hind limb paresis; 3 = hind limb paralysis; 4 = moribund; 5 = death.
After induction of EAE, in vivo MRI scans were made at different time points during the development of the disease: at day 9, animals have no clinical signs; at day 11, the first neurological aberrations occur and animals start to loose tail tonus; at day 14, the peak of the disease is reached; and at day 17, most animals have recovered. Controls were imaged at day 14. At each time point, BBB permeability and cellular infiltration were determined by Gd‐DTPA (0.5 mM; ©Magnevist, Shering, Germany) enhanced imaging (four animals per time point) and USPIO 7228 (600 µmol Fe/kg; Guerbet, Roissy CdG Cedex, France, Advanced Magnetics, Cambridge, MA, USA) enhanced imaging (six animals per time‐point).
Anaesthesia was induced with a mixture of fentanyl citrate (0.315 mg/kg), fluanisone (10 mg/kg) and midazolam (5 mg/kg), after which animals were prepared for mechanical ventilation by an endotracheal intubation. A tail vein was cannulated in animals subjected to a bolus of Gd‐DTPA. The animals were immobilized in a specially designed stereotactic device and placed in an animal cradle, which was inserted into the nuclear magnetic resonance (NMR) spectrometer. During the MRI experiments, the animals were ventilated with halothane (1%) in N2O/O2 (70/30). Expiratory CO2 was monitored, and the body temperature was maintained at 37°C with a heated water pad. An infrared sensor (Nonin Medical Inc., Plymouth, NH, USA) was attached to the animal hind paw to monitor heart rate and blood oxygen saturation.
High‐resolution MRI experiments were performed on a 4.7 T horizontal bore NMR spectrometer (Varian, Palo Alto, CA, USA), equipped with a high‐performance gradient insert (12 cm inner diameter; maximum gradient strength 220 mT/m). A Helmholtz volume coil (ϕ = 5 mm) and an inductively coupled surface coil (ϕ = 35 mm) were used for radio frequency transmission and signal reception, respectively. During the experiments, NMR gain was held constant. On a sagittal scout image, 35 contiguous coronal slices of 0.75 mm were defined covering the anterior part of the spinal chord and a large part of the brain. From these 35 slices, different MRI data sets were collected [FOV (field of view) = 3.5 × 3.5 cm2; matrix: 128 × 128; zero‐filled to: 256 × 256; in‐plane resolution: 273.5 µm2; two transitions] depending whether an animal was randomized to the Gd‐DTPA group or to the USPIO group.
From all animals subjected to BBB function measurement by Gd‐DTPA enhanced imaging, the following MRI data sets were collected:
(i) Quantitative T2 relaxation time maps: these maps were obtained by a mono‐exponential fit of four multi‐echo images. TR (repetition time) = 5000 ms; TE (echo time) = 20, 40, 60 and 80 ms. To overcome motion‐artifacts arising from ventricular pulsation, we triggered every 90° pulse with heart rate. The images with a TE of 60 ms were used for determination of region of interest (ROIs) (see also later).
(ii) Quantitative Gd‐DTPA enhanced T1‐weighted maps (GE‐T1W): these maps were calculated from two T1‐weighted spin echo images (TR = 650 ms; TE = 12.5 ms) before and after a bolus of 0.5 mmol/kg Gd‐DTPA [intravenously (i.v.), 8 min in circulation] with GE‐T1W = 100 × (T1W post Gd‐DTPA – T1Wpre Gd‐DTPA)/T1Wpre Gd‐DTPA. Pixel‐intensities thus display the percentage increase of the signal intensity of the T1‐weighted image due to Gd‐DTPA leakage.
Animals subjected to cellular infiltration studies received an i.v. bolus of USPIO 7228, under light ether anaesthesia, 24 h prior to the MRI experiments; this is the optimal time point (Weissleder et al., 1990; Dousset et al., 1999c). Pilot studies were performed using 300, 600 and 800 µM Fe/kg to determine the optimal dose of USPIOs to image EAE in animals 14 days after immunization (i.e. at the peak of the disease). From these pilot studies, 600 µM Fe/kg was selected as the optimal dosage. The USPIO agent 7228, an intravascular contrast agent, is composed of an iron core surrounded by an anionic dextran coat. The mean diameter of this particle (iron + dextran coating) is 30 nm, and its blood half‐life at the dosage of 600 µmol Fe/kg is ∼6 h in Lewis rats.
For these animals, only quantitative T2 relaxation time images were collected as described above.
MRI data evaluation
According to Paxinos and Watson (1986), the following representative ROIs were defined on T2‐weighted images (TE = 60 ms):
(i) Spinal cord: three slices posterior to the line spinal cord/cerebellum with spinal cord tissue;
(ii) Brain stem: all areas of the brain stem between the line spinal cord/cerebellum and cerebellum/cerebrum;
(iii) Cerebellum: the total area of the cerebellum;
(iv) Basal ganglia: all dorsal areas starting from the line cerebellum/cerebrum up to the utmost anterior part of the hippocampus;
(v) Cortical tissue of the cerebrum: all cortical areas starting from the line cerebellum/cerebrum up to the utmost anterior part of the hippocampus.
During the progression of EAE, no changes were observed in ventricle sizes; therefore, for all calculations, pixels arising from the ventricles were excluded by including only those pixels with T2 < 70 ms. As an objective measurement for the BBB integrity, we calculated the mean GE‐T1W value of these areas. For quantification of the USPIO burden, we defined in the USPIO‐loaded CFA control animals an area‐specific threshold in the T2‐maps, which was the mean T2 value minus 2 × SD. Thereafter the percentage pixels below each threshold for a specific area was calculated for each time point in the USPIO‐loaded EAE animals.
After MRI, animals were sacrificed. The brains were dissected, snap‐frozen in the vapour phase of liquid nitrogen and stored at –80°C. Cryostat sections (8 µm) were melted onto gelatin‐coated glass slides and dried in containers with silica gel. Slides were fixed with acetone (10 min) and were incubated with PBS with 10% foetal calf serum (Biowhittaker Europe, Verviers, Belgium) and 0.5% bovine serum albumin (Fraction V, Roche, Indianapolis, IN, USA) to prevent non‐specific binding. Thereafter, immunohistochemistry was performed. For detection of serum protein leakage into the brain parenchyma, sections were stained with a fluoroisothiocyanate (FITC) labelled polyclonal fibrinogen antibody (1:100; DAKO, Glostrup, Denmark). Infiltrated T‐cells were detected by monoclonal antibody (mAb) R7.3, directed against the T‐cell receptor, TCRαβ (Hunig et al., 1989) (biotinylated mIgG1; Pharmingen, San Diego, CA, USA). Macrophage markers mAb ED1 and mAb ED2 were used. Monoclonal antibody ED1 recognizes a lysosomal membrane related antigen expressed on both monocytes and macrophages, and mAb ED2 recognizes an antigen on a subset of mature rat macrophages including resident meningeal and perivascular macrophages in the CNS (Dijkstra et al. 1994) (biotinylated mIgG1; produced in the Department of Molecular Cell Biology, VUMC and commercially available via Serotec, Kidlington, Oxfordshire, UK). Slides were incubated with mAbs (2.5 µg/ml) for 1 h at room temperature. As conjugates alkaline phosphatase coupled streptavidin (Vector Laboratories, Burlingham, CA, USA) and streptavidin‐alexa‐488 and streptavidin‐alexa‐594 (both from Molecular Probes, Eugene, OR, USA) were used. Slides were rinsed with 0.1 M Tris pH 7.6. Alkaline phosphatase activity was demonstrated by incubation in AS‐BI substrate (Burstone, 1958) in 0.1 M Tris pH 8.7 for 10 min. Sections were rinsed, dried and mounted in Vectamount (Vector Laboratories). The presence of USPIOs was detected by Prussian blue staining of iron. The presence of fibrinogen in the CNS was visualized using a Nikon Eclipse 800 fluorescence microscope. The co‐localization of Prussian blue staining and ED1 expression was analysed using a 3I Marianas™ digital microscopy workstation (Intelligent Imaging Innovations, Denver, CO, USA).
Computerized analysis of immunohistochemistry
For quantification of the cellular infiltrates, brain sections at the level of 12.5 mm before Bregma, containing cerebellum and brainstem, of two animals per time point were examined with a microscope (Nikon Eclipse E800) and standardized recordings were made with a digital Nikon DXM1200 camera at ×10 objective. For each acquisition session, pictures were captured using a standardized procedure. To calculate the stained area, the digital image analysis program AnalySIS was used (Soft Imaging System GmbH, Münster, Germany). Each pixel in the photographs was divided into three‐colour components (hue, saturation and intensity). The threshold for these three colour components was defined in such a manner that only immunohistochemical positive areas were selected for analysis. The threshold was kept constant throughout the analysis of all animals. Ten pictures of the cerebellum were analysed per animal. The immunohistochemical positive area and intensity of positive pixels were defined for each picture. For analysis, the number of pixels in the immuno‐positive area was used. Data are expressed in arbitrary units.
Two animals from each group were perfused intracardially with 200 ml Sörensenbuffer (0.067M Na2HPO4.2H2O and 0.067 M KH2PO4, 4:1) followed by 300 ml of 3% paraformaldehyde/1.5% glutaraldehyde solution. To explore the uptake and localization of the USPIOs, tissue blocks from the upper part of the spinal cord, cerebellum, brainstem, basal ganglia, liver, spleen and cervical lymph nodes were fixed in 1.5% glutaraldehyde in PBS for several days at 4°C. After rinsing with PBS, the tissue blocks were osmificated (1 h in 2% OsO4 at 4°C), dehydrated, impregnated and embedded in a mixture of araldite and epoxyresin (Serva, Heidelberg, Germany).
Ultrathin sections were cut with an OMU III ultramicrotome (Leica, Mannheim, Germany), contrasted with uranyl acetate and lead citrate, and subsequently examined with a CM100 B10 twin Electron microscope (Philips, Eindhoven, The Netherlands).
Lovastatin treatment of EAE animals
To validate whether the effect of therapeutics can be monitored in vivo by USPIO‐enhanced T2imaging, four EAE animals were treated with lovastatin [4 mg/kg, intraperitoneally (i.p.); Merck Research Laboratory, Whitehouse Station, NJ, USA]. Treatment was started at day 0 and continued daily. At day 14, USPIO enhanced quantitative T2 images were obtained as described above.
Statistical analysis was performed by non‐parametric comparison using the Mann–Whitney U test. The relationship between the USPIO accumulation and the neurological aberrations or the cellular infiltration observed by immunohistochemistry were tested using the Spearman’s rank correlation coefficient. Statistical significance was defined as P < 0.05.
Clinical course of acute EAE
Acute EAE is a characterized by a monophasic course (Fig. 1). At day 9 post immunization, no clinical signs were observed in the immunized animals. At day 11, animals (16 out of 30) started to loose tail tonus and,at day 14, all animals (20 out of 20) were clinically ill (mean clinical score of 2.9 ± 0.4). Most animals (7 out of 10) were completely recovered at day 17. In the course of the disease, all animals experienced severe weight loss (Fig. 1). The CFA control animals did not show any clinical signs. For the MRI experiments, animals were imaged at day 9, 11, 14 and 17 after induction of EAE, while CFA control animals were imaged at day 14.
BBB permeability in the course of EAE
BBB integrity was quantified with Gd‐DTPA enhanced imaging. The presence of Gd‐DTPA in the brain parenchyma of different ROIs was determined at day 9, 11, 14 and 17 after induction of EAE. Fig. 2 shows typical examples of the macroscopic detection of Gd‐DTPA leakage by an increase of GE‐T1W values at different time points in animals suffering from EAE. At day 9 after induction of EAE, no difference was observed in GE‐T1W values compared with control animals (CFA control day 14). Interestingly, at the early onset of clinical signs (day 11; clinical score 0.5 ± 0.1), GE‐T1W values were increased in the spinal cord, brainstem and cerebellum as result of Gd‐DTPA leakage (Fig. 3). These values remained at this level at the peak of the disease (day 14) and returned to control values (spinal cord, brainstem, cerebellum) when the animals had recovered from the disease (day 17). No increase in GE‐T1W values was observed in the cortical area of the cerebrum.
USPIO accumulation in the CNS during EAE
EAE animals were injected with USPIOs (i.v.) 24 h prior to an MRI session. Quantitative T2 relaxation time images were measured at day 9, 11, 14 and 17 after induction of EAE. Macroscopic examination of the T2‐weighted images showed the appearance of hypointense abnormalities at day 11, mainly in the lower regions of the brain. At day 14 post immunization, these abnormalities occurred throughout the whole brain. Although less abundant, these abnormalities are also present at day 17 post immunization (Fig. 2). In control animals loaded with USPIOs, no abnormalities were observed in T2‐weighted images (data not shown). For quantification of the USPIO load in the different regions of the brain during the development of the disease, an area specific threshold was determined in USPIO‐loaded control animals, which was the mean pixel intensity of T2 value minus 2 × SD. The percentage of pixels below this area specific threshold was calculated for each USPIO‐loaded EAE animal (see Fig. 3). At day 9 after EAE induction, similar to the macroscopic analysis, no USPIO burden was observed in the different regions of interest (Fig. 3). At day 11, only a slight increase of USPIO burden was found. At the peak of the disease (day 14), USPIOs were present in all areas of the brain and especially in the brainstem, the cerebellum and the basal ganglia. No USPIOs were detected in the cortical area of the cerebrum (Fig. 3). The quantified USPIO burden in specific brain regions (i.e. the spinal cord, brainstem, cerebellum and basal ganglia) correlated significantly with the occurrence of neurological aberrations (Spearman correlation coefficient rspinal cord = 0.51, rbrainstem = 0.65, rcerebellum = 0.59 (all P < 0.01) and rbasalganglia = 0.41 (P < 0.05)).
Cellular infiltration and localization of USPIOs in the brain parenchyma
Brain sections were analysed immunohistochemically for the presence of fibrinogen (as a marker of BBB dysfunction) and cellular infiltrates. As early as day 11, fibrinogen was detected around most of the cerebral blood vessels, but hardly any cellular infiltrates could be detected (Fig. 4A, upper panel). At day 14, fibrinogen as well as cellular infiltrates (as stained by ED1, a monocyte marker) surrounded the cerebral blood vessels (Fig. 4A, lower panel), whereas in control animals, no fibrinogen leakage or cellular infiltrates could be detected (data not shown).
The presence of inflammatory cells in the course of EAE was analysed quantitatively in brain sections of the cerebellum and brain stem (Fig. 4B). In control and EAE animals at day 9, no cellular infiltrates were found. At day 11, few monocyte derived macrophages and T‐cells were present in the cerebellum and brainstem. At the peak of the disease (day 14), a more than ten‐fold increase of monocytes and T‐cells was observed, which substantially decreased at day 17. In contrast, ED2 antigen expression (which is selectively expressed on perivascular and meningeal macrophages) showed a two‐fold enhanced expression as early as day9 of the disease compared with controls. At the peak of the disease (day 14) and in the recovery phase (day 17), ED2 antigen expression was lower compared with day 9 of EAE (Fig. 4B).
Furthermore, the presence of infiltrated macrophages (ED1 positive) and T‐cells highly correlated with the accumulation of USPIOs in the cerebellum and brainstem (rcerebellum = 0.82 and rbrain stem = 0.8; P < 0.01). Similar to the USPIOs, the presence of T‐cells and ED1 positive monocytes also correlated with the neurological score of the animals (both r = 0.90, P < 0.01). The expression of the ED2 antigen did not correlate with the presence of USPIOs or neurological score (rUSPIO = –0.42, rneurological score = 0.082, both P > 0.05).
To establish cellular uptake of USPIOs, brain sections were stained with Prussian blue, which specifically reacts with iron. At the peak of the disease (day 14), affected areas showed focal intracellular spots in perivascular infiltrates (Fig. 5A). In USPIO‐loaded CFA control animals, no Prussian blue positivity was observed (data not shown). Immunofluorescence microscopy revealed a clear co‐localization of ED1 positive monocytes with Prussian blue, indicating that a proportion of infiltrated monocytes contain USPIOs (Fig. 5A). The localization of USPIOs in perivascular infiltrates was further confirmed by electron microscopy (Fig. 5B, indicated by arrows). In peripheral tissues (spleen and liver), USPIOs were also present in phagocytic cells (data not shown).
Lovastatin treatment reduces clinical signs and USPIO burden
To test whether USPIOs can be used to monitor the effect of (new) therapeutics, EAE animals were treated daily with the immunomodulator lovastatin (4 mg/kg i.p.), which has been suggested to reduce cellular infiltrates. Lovastatin treatment, which was started at day 0, significantly reduced the clinical scores of EAE animals, from 2.9 ± 0.4 to 0.9 ± 0.5 at day 14 (Fig. 6A). This effect was also observed with USPIO enhanced T2 imaging in the brainstem, cerebellum and basal ganglia of lovastatin‐treated animals. Hardly any abnormalities were observed on the quantitative T2 images, which resulted in a significant reduction of the percentage pixels with a decreased T2‐value (Fig. 6B). Immunohistochemical analysis indicated that lovastatin inhibited the infiltration of ED1 positive monocytes and T‐cells into the brain in EAE animals (data not shown).
With this study, we quantitatively visualized cellular infiltration and increased BBB permeability as distinct neuropathological processes using USPIOs and Gd‐DTPA contrast MRI in EAE animals. Moreover, we showed that MRI of USPIO load is a powerful tool to detect monocyte infiltrates and to evaluate the therapeutic effects of administered compounds.
In our EAE animals, increased BBB permeability as imaged by GdDTPA‐enhanced MRI marked the onset of the disease and was already at a maximal level when animals started to develop first clinical signs, but did not correlate with the presence of cellular infiltrates. Enhanced BBB permeability persisted on day 14 (peak of the disease) and started to decline upon recovery of the animals at day 17. Immunohistochemical analysis confirmed the presence of blood‐borne compounds such as fibrinogen in the brain parenchyma at the start of the disease. The cortical tissue remains free from leakage of BBB, which is in accordance with studies on vascular leakage of dyes in this model (Juhler et al., 1984, 1985). General Gd‐DTPA enhancement has been related only to the onset of clinical signs in EAE models, but no correlations have been made with the presence of cellular infiltrates in the different brain regions (Hawkins et al., 1990; Namer et al., 1992; Karlik et al. 1993; Seeldrayers et al., 1993; Morrissey et al., 1996).
So far, it is unknown what causes the opening of the BBB early in the course of EAE. The increased permeability may be the result of the activation of the perivascular macrophages. In our study, we observed an increased expression of ED2 (an activation marker) on perivascular and meningeal macrophages before the onset of clinical signs of EAE; this is in agreement with previous findings (Polfliet et al., 2002). Considering their early activation, it is likely that, by secreting inflammatory mediators (Bauer et al., 1993; Mato et al., 1996; Prat et al., 2001), these phagocytic cells play an important role in the opening of the BBB—as observed in our study. Together these data implicate that monitoring of BBB permeability by Gd‐DTPA enhancement is a reliable marker for early inflammatory responses at the level of the BBB, but not for the presence of cellular infiltrates.
Non‐invasive imaging of the occurrence of cellular infiltrates, especially active monocyte‐derived macrophages, may provide valuable information about the pathophysiological processes in the CNS that cannot be detected using Gd‐DTPA imaging. Therefore, we used USPIOs as a contrast agent to visualize cellular infiltration in vivo. We are the first to quantitatively describe the correlation between the presence of infiltrated macrophages in the CNS with MRI of USPIO load in the course of EAE. USPIO accumulation in the brain, as quantified by the percentage area with abnormally short T2 relaxation times, correlated both with clinical disease scores and cellular infiltrates of monocyte derived macrophages and T‐cells, but not with perivascular macrophages (as determined by ED2 antigen expression). In a subgroup of three animals, a second MRI‐experiment was performed 3 days after the initial MRI scan to follow the dynamics of USPIO load in the brain (data not shown). These experiments showed that, independent of an increase or decrease in neurological score, the USPIO load almost completely regressed in the observed areas indicating a high turnover of macrophage influx.
It is still unclear by what mechanism USPIOs enter the brain parenchyma. It is suggested that USPIOs are incorporated in mononuclear (phagocytic) cells in the bloodstream, which migrate towards the CNS under inflammatory conditions (Rausch et al., 2002). Alternatively, USPIOs may possibly diffuse across a ruptured BBB (Muldoon et al., 1999) or cross the brain endothelium by means of transcytosis (Weissleder et al., 1990), and are subsequently taken up by residential macrophages.
We show here that the USPIOs in the CNS are present in newly infiltrated monocytes. Immunohistochemical analysis revealed that the iron particles are specifically localized within newly infiltrated ED1 positive cells, but not in ED2 positive perivascular macrophages. Further analysis by electron microscopy clearly located iron particles within lysosomes of these newly infiltrated cells. Moreover, we observed that the quantitative USPIO‐load was coincident with the quantitative ED1 immunohistochemistry. These observations suggest that USPIOs are either taken up in newly infiltrated monocytes after their leakage across the impaired BBB or that USPIO loaded monocytes have migrated across the brain endothelium. Although disruption of the BBB was maximal at day 11 (as shown by Gd‐DTPA enhancement and fibrinogen presence), we only observed the presence of a few USPIOs at this time point, suggesting that there is no passive passage of USPIOs across the damaged BBB. We therefore postulate that cerebral USPIO accumulation is mainly the result of the extravasation of USPIO‐loaded monocytes from the bloodstream into the brain.
We have confirmed the uptake of USPIOs by monocytes (as described by Moore et al., 1997). In vitro experiments were performed with freshly isolated rat monocytes loaded with USPIOs in a concentration comparable to the in vivo situation. These USPIO‐loaded monocytes have enhanced granulation—as observed by fluorescence activated cell sorting (FACS) analysis—indicating the uptake of iron, which was confirmed by Prussian blue staining and electron microscopy. After uptake of USPIOs, the monocytes remained viable (as determined by FACS analysis and Trypan blue exclusion) and did not show differences in their migration across brain endothelium compared with vehicle treated monocytes. Furthermore, no increase in the production of reactive oxygen species as a marker of monocyte activation was observed, indicating that the uptake of USPIOs by monocytes does not lead to activation of the cells (unpublished data).
We here quantitatively demonstrate that increased permeability of the BBB (as determined by Gd‐DTPA enhanced imaging) occurs prior to monocyte infiltration (as determined by USPIO enhanced imaging) in specified regions of the CNS. Previously, Dousset et al. (1999a) compared the application of both these contrast agents in acute EAE at the peak of the disease. In contrast to our findings, no Gd‐DTPA enhancement (using a similar Gd‐DPTA dosage) was found in clinically severely ill rats. The advantage of USPIOs as a contrast agent over Gd‐DTPA is that the actual lesion areas become significantly enlarged on MRI scans, revealing even small lesions (Xu et al., 1998). One of the strategies of newly developed therapies for multiple sclerosis is to interfere in the process of cellular infiltration, thereby limiting new lesion formation. To evaluate the beneficial effect of such treatments, in vivo monitoring of monocyte infiltration and subsequent macrophage activity in time, is an essential tool. Here we show that lovastatin, acting as an immunomodulator, significantly ameliorates clinical signs in EAE as reported previously (Stanislaus et al., 1999, 2001). This reduction of clinical scores correlated highly with the lack of abnormalities on T2‐relaxation time images at the peak of the disease compared with untreated EAE animals. This suggests that lovastatin treatment of EAE animals reduces cellular infiltration into the CNS. Thus, the use of MRI contrast agents based on USPIOs is suggested to be a sensitive method to evaluate the efficacy of new therapeutics in EAE and, possibly, multiple sclerosis. It has been described that USPIOs are well tolerated in humans and USPIOs have already been applied to the assessment of lymph node metastases (Anzai et al., 1994; Harisinghani et al., 1997; Sigal et al., 2002).
In conclusion, we have quantitatively described that, in the development of inflammatory lesions in EAE, BBB leakage and cellular infiltration can be distinguished in vivo by Gd‐DTPA and USPIO contrast MRI. Furthermore, we have provided evidence that MRI of USPIO load is a valuable tool to assess the effect of lovastatin as an example of a promising treatment strategy to prevent infiltration of monocyte‐derived macrophages into the CNS, hence limiting disease progression.
We wish to thank Dr R.J.P. Musters (Laboratory for Physiology, VUMC, Amsterdam, The Netherlands) for acquiring the images with the 3I Marianas™ digital microscopy workstation. This work was supported by grants from Stichting Multiple Sclerosis Research, The Netherlands and by NWO Medische Wetenschappen, The Netherlands.