Magnetic resonance imaging of blood brain/nerve barrier dysfunction and leukocyte infiltration: closely related or discordant?

Abstract

Unlike other organs the nervous system is secluded from the rest of the organism by the blood brain barrier (BBB) or blood nerve barrier (BNB) preventing passive influx of fluids from the circulation. Similarly, leukocyte entry to the nervous system is tightly controlled. Breakdown of these barriers and cellular inflammation are hallmarks of inflammatory as well as ischemic neurological diseases and thus represent potential therapeutic targets. The spatiotemporal relationship between BBB/BNB disruption and leukocyte infiltration has been a matter of debate. We here review contrast-enhanced magnetic resonance imaging (MRI) as a non-invasive tool to depict barrier dysfunction and its relation to macrophage infiltration in the central and peripheral nervous system under pathological conditions. Novel experimental contrast agents like Gadofluorine M (Gf) allow more sensitive assessment of BBB dysfunction than conventional Gadolinium (Gd)-DTPA enhanced MRI. In addition, Gf facilitates visualization of functional and transient alterations of the BBB remote from lesions. Cellular contrast agents such as superparamagnetic iron oxide particles (SPIO) and perfluorocarbons enable assessment of leukocyte (mainly macrophage) infiltration by MR technology. Combined use of these MR contrast agents disclosed that leukocytes can enter the nervous system independent from a disturbance of the BBB, and vice versa, a dysfunctional BBB/BNB by itself is not sufficient to attract inflammatory cells from the circulation. We will illustrate these basic imaging findings in animal models of multiple sclerosis, cerebral ischemia, and traumatic nerve injury and review corresponding findings in patients.

Keywords: blood brain barrier; contrast-enhanced MRI; gadofluorine; gadolinium-DTPA; iron oxide nanoparticles; neuroinflammation.

Figure 1

Gf enhances detection of inflammatory lesions with a dysfunctional BBB in EAE. A coronal T1-w image of the brain 24 h after systemic application of Gf to a diseased rat displays bright enhancement in the periventricular region (A). By contrast, corresponding coronal T2-w MRI, the standard measure for overall disease burden in MS, shows only a tiny periventricular lesion [arrow in (B)]. Sagittal T1-w MRI reveals, amongst other contrast enhancing lesions, a large Gf positive lesion in the dorsal column of the cervical spinal cord [arrow in (C)] that is verified on a macroscopic preparation of the entire medulla (D). Foci with Gf deposition appear pink due to the coupling with a carbocyanine dye. Intraindividual comparison of Gf and Gd-DTPA enhancement on coronal slices shows extensive Gf enhancement in the left posterolateral cervical spinal cord (F) while Gd-DTPA uptake is far more limited (E). On corresponding T2-w MRI the lesion is indiscernible (G). Adapted from Bendszus et al. (2008).

Figure 2

Gf improves EAE lesion detection in optic nerves (ON) belonging to the CNS (A-D). Twenty-four hours after systemic administration of Gf coronal T1-w MRI shows bilateral disturbance of the blood-optic nerve barrier [arrows in (A)]. The macroscopic preparation of the left ON (B) confirms extravasation of labeled Gf. Thereafter the ON was embedded in paraffin and longitudinal sections were stained for macrophages (ED-1) (C) and myelin (Luxol fast blue) (D). Note that the Gf enhancing ON is heavily infiltrated by ED-1 positive macrophages (C) and almost completely demyelinated [lack of deep blue staining in (D)]. (E,F) depict the spatial discrepancy between cellular inflammation and BBB dysfunction in EAE. Coronal CISS (constructive interference steady state) MRI of a severely affected rat 24 h after SPIO application shows strong signal loss in both optic nerves [arrows in (E)]. By contrast, the same animal lacks retrobulbar Gf enhancement on corresponding T1-w MRI 24 h after Gf administration (F) indicating that macrophages infiltrated the ON without concomitant disturbance of the blood-optic nerve barrier. Since acute macrophage infiltration as shown in (E) does not immediately cause leakage of the BBB allowing access of Gf (F) we speculate that BBB dysfunction is a secondary and delayed event dependent on previous macrophage infiltration. Adapted from Bendszus et al. (2008) (A) and Ladewig et al. (2009) (E,F).

Figure 3

Visualization of transient BBB opening remote from a photothrombotic lesion by Evans Blue extravasation (A,B), Gd-DTPA (C,E) and Gf enhanced MRI (D,F). Rats received Evans Blue i. v. immediately after PT induction or with a delay of 48 h (for 72 h analysis) and were sacrificed at given time points. Total brain preparations (A) and coronal brain sections (B) of the animals show that Evans Blue extravasation starts at the lesion site, but extends to the remote ipsilateral cortex and the corpus callosum within the first day. Coronal brain slices at the level of the lesion are always shown on top, the other six in the row represent subsequent 1 mm slices located frontally to the lesion (B). At day 3 breakdown of the BBB indicated by Evans blue extravasation is restricted to the PT lesion. (E) shows Gd-DTPA enhancement of the photothrombotic lesion within the first 24 h. Note that, in contrast, Gf enhancement occurs within in the lesion, but also affects the ipsilateral cortex and the corpus callosum spared by Gd-DTPA (F). Moreover, T1-w MR images anterior to the PT lesion exhibit no Gd-DTPA (C), but strong Gf enhancement (D) according to Evans blue extravasation shown in (A,B). Reproduced from Stoll et al. (2009b).

Figure 4

Visualization of breakdown of the blood nerve barrier (BNB) (A,B) and inflammation in experimental nerve crush (C–F). Coronal images depict the pelvis and both thighs of a rat lying in prone position with both legs positioned in a round surface coil (CISS sequence; slice thickness 1 mm). Note that Gf accumulates in the degenerating distal stump on T1-w MRI [arrow in (A)] and binds to peripheral nerve structures as revealed by fluorescence of carbocyanine-labeled Gf (B). Gf enhancement ceases not until successful regeneration (not shown). By contrast to breakdown of the BNB, macrophage infiltration is restricted to the early phase of Wallerian degeneration. Five days after sciatic nerve crush focal signal loss is present at the lesion site and distally due to the invasion of SPIO-labeled macrophages from the blood (C). The corresponding paraffin section stained for iron confirms the infiltration of numerous iron-laden macrophages in the degenerating nerve segment (D). At day 8, macrophage infiltration is restricted again to the lesion site and ceases thereafter (E). Correspondingly, distal nerve segments no longer show iron-positive cells after application of SPIO as shown for day 10 in (F). The BNB, however, is still leaky at that time (not shown) indicating that macrophage infiltration occurs within a narrow time window and persistent BNB disturbance does not per se induce cellular infiltration. Adapted from Bendszus et al. (2005b); (A,B) and Bendszus and Stoll (2003) (C–F).