SPIO/USPIO enhanced MRI
Gadolinium-DTPA and Gf enhanced MRI are crude indicators of BBB or BNB disturbances, but they do not specifically visualize cellular immune responses. Inflammation, however, plays a major role in disorders of the nervous system. While under healthy conditions immune cell access to the nervous system is restricted, leukocytes readily cross the BBB or BNB in the context of pathophysiological processes and enter the neural tissue guided by cell adhesion molecules and chemokines (Man et al., 2007).
There is an ongoing controversy whether immune cell invasion is linked to BBB disruption or occurs independently (see below). A prerequisite for solving this issue are imaging tools that allow monitoring of cell migration in vivo. From a clinical perspective, tracking of inflammatory cells could moreover help to identify active phases of CNS and PNS inflammation and to monitor the efficacy of therapeutic interventions. Thus, it is not surprising that cellular neuroimaging has become a field of intense research effort. The preferentially applied contrast agents for cellular MRI are iron oxide nanoparticles and lately, PFC (see Perfluorocarbon Enhanced 19F MRI). Depending on their hydrodynamic particle size iron nanoparticles are classified into superparamagnetic iron oxide particles (SPIO; 50–200 nm diameter) and their ultra and very small variants (USPIO; ∼35 nm diameter; VSOP < 10 nm diameter). They all possess relatively large negative magnetic susceptibilities, featuring a more extensive shortening of T1 and T2 relaxation times than Gd-DTPA. Hence, their sensitivity is much higher than that of Gd compounds. The easiest and safest method of iron labeling is spontaneous uptake by blood-borne cells after systemic contrast agent application (Bulte, 2009). If thereafter labeled cells are attracted to a target organ they can be visualized by MRI. Importantly, the property to phagocytose and to migrate to sites of inflammation turns macrophages into ideal targets for cellular MRI. Apart from macrophage phagocytosis SPIO/USPIO are cleared from the circulation by cells of the reticuloendothelial system (RES) in liver and spleen. The extent of cellular labeling in relation to clearance by the RES depends on iron oxide particle size and the net charge of the polymer coating (Hoehn et al., 2007). Safety concerns for both, SPIO and USPIO compounds are little. A recent meta-analysis of 37 phase I to III clinical trials with ferumoxtran-10, an USPIO agent, revealed back pain, pruritus, headache, and urticaria as the most frequent adverse events (Bernd et al., 2009). They are usually mild and of short duration (Anzai et al., 2003). Similarly, SPIO have been associated with low human toxicity. In numerous clinical trials they were shown to cause neither cardiovascular side effects nor clinically relevant laboratory changes (Bluemke et al., 2003; Reimer and Balzer, 2003).
EAE and MS
Several proof-of-principle studies have demonstrated that USPIO and SPIO enhanced MRI allow visualization of macrophage infiltration in EAE (Dousset et al., 1999; Rausch et al., 2003; Floris et al., 2004). Brochet et al. (2006) showed that rats with USPIO positive lesions at the first attack suffer from more severe clinical affection and extensive axonal damage at the second EAE bout. This implies that the extent of macrophage infiltration in early EAE may predict the ensuing disease course. Interestingly, in a follow-up study by the same group severely affected rats with USPIO enhancement at the onset of disease showed an imbalanced, proinflammatory macrophage activation profile, both in CNS lesions and the peripheral blood (Mikita et al., 2011). Moreover, macrophage invasion spread from the upper spinal cord and brainstem to the cerebellum and subcortical regions during disease progression. When USPIO were injected in the recovery phase no signal abnormalities were observed indicating that macrophage infiltration in EAE is a timely restricted event that clearly depends on the disease stage. In addition, iron-enhanced MRI was applied as a non-invasive tool for preclinical treatment surveillance. Deloire et al. (2004) demonstrated that macrophage recruitment to the EAE lesions detected by USPIO enhanced MRI is not fully blocked under therapy with natalizumab, a VLA4-antagonist that is in frequent use in patients with relapsing-remitting MS. Moreover, the efficacy of Fingolimod, an oral drug that was recently approved for MS treatment, was successfully monitored by USPIO enhanced MRI in EAE (Rausch et al., 2004).
In MS it is a common conception that the poor clinico-radiological association may be explained in part by diffuse inflammatory activity in the so-called normal appearing white matter (NAWM) which is concealed behind an intact or repaired BBB (Barkhof, 2002; Kutzelnigg et al., 2005). Indeed, microscopic inflammation within the NAWM not amenable by conventional MRI seems to contribute more strongly to disability than the T2 lesion load (Parry et al., 2002; Traboulsee et al., 2003). Thus, it was not surprising that the application of iron-enhanced MRI in EAE and the transfer of this technology to MS patients in small trials have provided insights into CNS inflammation that exceed conventional Gd-DTPA enhanced MRI. Vellinga et al. (2008) detected 188 USPIO positive cerebral lesions in 14 patients with active relapsing-remitting MS. Interestingly, the vast majority of USPIO positive lesions (144/188) showed no concomitant Gd-DTPA enhancement. Furthermore, none of the three different types of USPIO enhancement (i) focal lesions (ii) “return to isointensity” lesions and (iii) ring-enhancing lesions was particularly related to Gd-DTPA uptake (Vellinga et al., 2008). A more recent clinical trial by the same group suggested USPIO as a potential marker for diffuse inflammation in the NAWM of relapsing-remitting and primary progressive MS patients (Vellinga et al., 2009). T1 histogram and region-of-interest analysis in the NAWM showed diffuse T1-shortening after USPIO injection in 16 MS patients indicative of subtle inflammatory activity, but not in gender- and age-matched healthy controls.
Ischemic stroke
It has long been established that cerebral ischemia induces a profound inflammatory response involving neutrophils, T-cells, and macrophages (Stoll et al., 1998). In the simple model of photothrombotic infarction (see above), entry of phagocytic hematogenous macrophages is delayed by several days and peaks around days 6–9 after lesion induction (Schroeter et al., 1997). Accordingly, SPIO-induced signal alterations did not occur until day 5 despite persistent breakdown of the BBB in PT lesions. However, at day 6 a hypointense rim appears followed by signal loss in more central areas reflecting the influx of hematogenous macrophages as confirmed by histological analysis (Kleinschnitz et al., 2005). Macrophage inflammation could also be visualized by SPIO/USPIO enhanced MRI after transient or permanent middle cerebral artery occlusion (MCAO), but results are more heterogeneous and conflicting. Some groups found contrast-induced signal loss in the lesion boundary 24 h after permanent MCAO (Rausch et al., 2001). In a transient MCAO model, Denes et al. (2007) did not find USPIO-related signal loss within the first 3 days after stroke induction. Moreover, neither focal signal intensity changes nor iron-positive macrophages were detected in the ischemic hemisphere of Wistar rats when USPIO were applied at a subacute stage 6 days after MCAO (Farr et al., 2011). By contrast, Kim et al. (2008) found areas of signal loss in SPIO enhanced MRI 3 days after reperfusion that corresponded to the accumulation of iron-laden macrophages. A possible explanation for these contradictory results is that the labeling efficacy of monocytes by USPIO in comparison to larger nanoparticles such as SPIO is rather low (Oude Engberink et al., 2007). Moreover, the timing of the contrast agent application and the subsequent MRI appears to have critical impact on the processes depicted by iron-enhanced MRI. Hence, especially in the early infarct phase, focalized signal alterations might rather be caused by trapping of the iron particles in the vasculature than phagocyte infiltration (Desestret et al., 2009).
Several open-label pilot trials with USPIO enhanced MRI were conducted in patients with ischemic stroke. Saleh et al. (2004) applied USPIO in a series of 10 patients 5–6 days after stroke onset. MRI scans were performed 24 and 48 h after USPIO infusion. As principal finding two distinct USPIO-related signal changes were observed: blood pool effects that appeared as signal loss on T2/T2*-w images and decreased from the first to the second scan as well as parenchymal contrast enhancement on T1-w images that increased over time and was attributed to macrophage infiltration. When applied early (24–36 h after stroke) in a subsequent study, USPIO enhancement was spatially heterogeneous and only present in a minority of patients (Saleh et al., 2007). Nighoghossian et al. (2007) recruited patients with anterior circulation stroke and administered USPIO 6 days after admission. Three days later, 9 of 10 patients showed USPIO enhancement in the brain parenchyma. Interestingly, whilst most patients featured mild Gd-DTPA uptake, the patient with the most severe BBB breakdown did not exhibit USPIO enhancement.
Nerve trauma
Traumatic injury to peripheral nerves similarly induces a profound inflammatory response with macrophage infiltration (Stoll et al., 1989) that leads to rapid removal of myelin debris and a growth-promoting cellular and molecular milieu. Thus, WD is the prototype of a tissue protective M2 type macrophage response (Ydens et al., 2012). SPIO enhanced MRI helped to define the kinetics of macrophage entry into the degenerating nerve segments which starts around day 1 or 2 at the lesion site and extends to the entire distal stump within the first 2 weeks after injury (Bendszus and Stoll, 2003; Figures 4C,D). Accumulation of SPIO-laden macrophages was visible as focal signal loss on T2-w MRI. When SPIO particles were applied 10 days after crush or later degenerating nerves did no longer exhibit signal loss (Figure 4E) despite the presence of numerous myelin-laden macrophages in the endoneurium. There are two possible explanations for this finding: (i) SPIO-based MRI depicts active migration of macrophages from the circulation into nerves, and lack of signal at later stages of WD indicates no further macrophage recruitment (Figure 4F). (ii) Alternatively, myelin-loaded macrophages no longer phagocytose iron particles within the injured nerves. We favor the first possibility since the dynamics of macrophage infiltration shown by SPIO enhanced MRI closely resembles the local expression pattern of macrophage attracting chemokines, which ceases around day 10 (Toews et al., 1998; Tofaris et al., 2002). Moreover, our data suggests that simple leakage of SPIO particles through a defective BNB does not significantly change the intrinsic nerve MR signal, at least in the PNS. These results were recently confirmed in a rodent model of radicular pain. Seven days after transient dorsal root compression T2*–w MRI showed significant iron-induced signal alterations in the nerve roots of the injured, but not of the sham-operated group after systemic application of SPIO (Thorek et al., 2011).
Perfluorocarbon enhanced <sup>19</sup>F MRI
Despite its excellent sensitivity iron particle based cellular MRI has several shortcomings. These include false-positive results caused by hemorrhages, blood pool effects and, in some instances, (depending on the half-time of the compound) passive diffusion via a defective BBB or BNB. Additionally, the large 1H background signal from mobile water renders unambiguous detection of labeled cells in vivo difficult especially if their biodistribution is unclear. Recently, fluorine (19F) MRI has emerged as an alternative approach for cellular imaging (Stoll et al., 2012; Temme et al., 2012). 19F markers exhibit favorable MR imaging characteristics, such as a magnetic sensitivity close to the proton nucleus and high natural abundance. Moreover, because of the lack of endogenous 19F-containing molecules in the body signals originating from injected 19F compounds are specific. Another virtue of 19F MRI is that it can be performed in a quantitative manner. Ideal fluorine tracers should provide a high payload of 19F nuclei. PFC compounds fulfill this requirement and thus were established as tracers for 19F MRI in recent years. In animal studies PFC agents were generally well-tolerated (Ebner et al., 2010). Several PFC compounds have also been elaborately studied as artificial blood substitutes (Spahn et al., 2002). In one of these clinical trials cerebral hemorrhages occurred after cardiopulmonary bypass surgery but thorough analysis of the safety data revealed that the study conduct, and not the PFC emulsion itself was responsible for the adverse events (Riess, 2006). Moreover, extensive analysis disclosed no perturbation of hemostasis or blood viscosity after i. v. treatment with PFC that could be related to the observed bleeding tendency (Riess, 2006).
Flogel et al. (2008) were the first to show that systemically injected 19F emulsions are efficiently phagocytosed by circulating monocytic cells. Moreover, in mice with myocardial infarction they were able to monitor a time-dependent PFC accumulation in the infarct area. It is well known that myocardial ischemia induces an inflammatory response dominated by cells of the monocyte/macrophage system. Consistently, rhodamine-labeled PFC allowed the identification of 19F-positive cells within the infarction as macrophages. Subsequently, in vivo 19F MRI was successfully applied to monitor immune cell responses in mice with LPS-induced pneumonia (Ebner et al., 2010), abscess formation (Hertlein et al., 2011), and in models of acute allograft rejection (Flogel et al., 2011; Hitchens et al., 2011). However, studies using 19F MRI in the nervous system are sparse. Probably due to the low accumulation of PFC compounds in inflammatory lesions of the CNS/PNS compared to other organ systems sensitivity is a major concern (Stoll et al., 2012). Lately, we established 19F MRI to depict macrophage infiltration in a rat model of focal inflammatory peripheral nerve injury (Weise et al., 2011). Focal injection of lysolecithin chemically dissolves myelin sheaths and, thereby, elicits a strong inflammatory response within the demyelinated nerve segment (Griffin et al., 1990). In vivo MRI 5 days after sciatic nerve damage revealed massive migration of 19F labeled cells to the injured nerve section. Intraneural application of saline to the contralateral nerve provoked a slight inflammatory reaction restricted to the perineurium which could also be visualized by 19F MRI. However, quantification of the signal strength by ex vivo 19F spectroscopy indicated a significantly higher number of fluorine-labeled cells in the lysolecithin-damaged nerve than in the contralateral control. Attempts to visualize macrophage responses in EAE by 19F MRI in vivo failed so far, most likely due to insufficient sensitivity (reviewed in Stoll et al., 2012). Thus, despite its superiority in specificity over SPIO/USPIO-based MRI, better coils and more sensitive MRI sequences are needed for further in vivo application of 19F MRI in the nervous system.