Iron is a sensitive biomarker for inflammation in multiple sclerosis lesions

Abstract

MRI phase imaging in multiple sclerosis (MS) patients and in autopsy tissue have demonstrated the presence of iron depositions in white matter lesions. The accumulation of iron in some but not all lesions suggests a specific, potentially disease-relevant process, however; its pathophysiological significance remains unknown. Here, we explore the role of lesional iron in multiple sclerosis using multiple approaches: immunohistochemical examination of autoptic MS tissue, an in vitro model of iron-uptake in human cultured macrophages and ultra-highfield phase imaging of highly active and of secondary progressive MS patients. Using Perls' stain and immunohistochemistry, iron was detected in MS tissue sections predominantly in non-phagocytosing macrophages/microglia at the edge of established, demyelinated lesions. Moreover, iron-containing macrophages but not myelin-laden macrophages expressed markers of proinflammatory (M1) polarization. Similarly, in human macrophage cultures, iron was preferentially taken up by non-phagocytosing, M1-polarized macrophages and induced M1 (super) polarization. Iron uptake was minimal in myelin-laden macrophages and active myelin phagocytosis led to depletion of intracellular iron. Finally, we demonstrated in MS patients using GRE phase imaging with ultra-highfield MRI that phase hypointense lesions were significantly more prevalent in patients with active relapsing than with secondary progressive MS. Taken together, our data provide a basis to interpret iron-sensitive GRE phase imaging in MS patients: iron is present in non-phagocytosing, M1-polarized microglia/macrophages at the rim of chronic active white matter demyelinating lesions. Phase imaging may therefore visualize specific, chronic proinflammatory activity in established MS lesions and thus provide important clinical information on disease status and treatment efficacy in MS patients.

Conflict of interest statement

Competing Interests: DP has received funding from Five Prime, Biogen Idec, Novartis and Teva Pharmaceuticals for investigator-initiated trials. He also received consulting fees from Biogen Idec. AB has received research grants from Biogen Idec, Novartis, Actillion, Accorda and MerckSerono. He has also received consulting fees from Biogen Idec, Genzyme, Teva Neuroscience, Novartis, MerckSerono and Metronics. This does not alter the authors' adherence to all the PLOS ONE policies on sharing data and materials.

Figures

Figure 1. Iron deposition in white matter MS plaques.

The overview image [a] shows an oil-red O-stained section with an actively demyelinating lesion and a demyelinated lesion . The actively demyelinating lesion (1) contains an abundance of large, myelin-laden macrophages as indicated by the presence of oil-red O positive material within CD68+ macrophages [b, c]. These macrophages do not contain iron as indicated by Perls' staining [d], or the iron storage protein ferritin [e] and express the M2 polarization marker CD206 [g] but not iNOS [f]. In the demyelinated lesion , at least two populations of macrophages were observed: in the lesion center , small round CD68+ macrophages contain condensed myelin [h, i]. Occasional macrophages contain iron [j] and correspondingly, ferritin [k] and express CD206 [m] but not iNOS [l]. In contrast, macrophages at the lesion rim (3) have a ramified appearance and do not contain oil-red O positive material [n, o]. These cells contain large amounts of iron [p] and iron-storing ferritin [q] and are iNOS positive [r] but CD206 negative [s], suggesting M1 polarization.

Figure 2. Iron and myelin uptake in human monocyte-derived macrophages (MDMs).

Non-polarized M0, M1- and M2-polarized macrophages were stained for neutral lipids (oil red-O) and iron (Perls). In untreated cultures, macrophages did not contain iron or significant amounts of lipids [a–c]. In macrophages exposed to FeCl3 (10 µM; 10 hrs), Perls' staining showed polarization-dependent iron uptake [d–f]. Exposure of iron-rich macrophages to purified human myelin (30 µg/ml; 24 hrs) lead to depletion of iron from myelin-phagocytosing macrophages in all polarization states [g–i]. Presence of iron within macrophages was closely mirrored by macrophage expression of ferritin: macrophages without iron-load express little ferritin [aa–cc], while iron loading leads to increased expression of ferritin [dd–ff], particularly in M1 macrophages [ee]. Addition of myelin to iron-rich macrophages resulted in substantial reduction of ferritin in all polarization states [gg–ii]. In macrophages exposed first to myelin [j–l] and subsequently to iron, iron accumulation was significantly reduced compared to that in naive macrophages [m–o]. When M1 macrophages were incubated with myelin in the presence of the receptor-associated protein (GST-RAP), which binds to the presumed receptor for myelin ingestion, LRP-1, and inhibits interaction with other ligands, myelin ingestion was prevented [p]. Blockade of myelin ingestion in M1 macrophages with GST-RAP and subsequent exposure to FeCl3 led to unimpeded iron uptake [q]. In contrast to myelin, internalization of fluorescent polystyrene microspheres by M1 macrophages did not prevent subsequent iron uptake [r].

Figure 3. Myelin decreases 55Fe uptake by MDMs.

Human MDMs were seeded in 96 well dishes at a cell density of 8×104 cells/well and polarized as described in Methods. Incubation with increasing concentrations of 55FeCl3 and non-radioactive FeCl3 resulted in dose-dependent and polarization-dependent iron uptake (M1>M2>M0). When human myelin was added (30 µg/ml) 24 hrs prior to exposure with iron, iron uptake was significantly reduced in all polarization states. Adding fluorescent-labeled polystyrene microspheres (1 ug/ml) instead of myelin did not significantly reduce radioactive iron uptake. In separate experiments, the iron chelator desferoxamine (100 µM) and hepcidin, which prevents cellular iron export by binding to the iron channel ferroportin (700 nM), were added as controls. Cell-associated radioactivity was determined in triplicate samples. The mean value of three separate experiments (± SEM) are shown. Statistical significance was obtained by comparing the values between cells incubated with and without myelin. *p<0.05; ** p<0.01.

Figure 4. Iron uptake enhances M1 polarization.

A. Human MDMs were seeded in 48 well dishes at cell density of 2×106 cells/ml and polarized according to protocol. Various concentrations of FeCl3 (0–100 µM) were added to the differentially polarized macrophages. Where indicated, cells were pre-incubated with human myelin (30 µg/ml) for 24 hrs before addition of FeCl3. After 10 hrs, cell-free supernatants were collected and TNF-α and IL-10 concentrations were measured by ELISA. Iron uptake increased secretion of TNF-α in M1-polarized but not in M0 and M2-polarized macrophages and reduced IL-10 in M2-polarized macrophages. Cytokine secretion did not change in myelin-laden macrophages after exposure to iron. Results are expressed as means ± SEM from three separate experiments. *p<0.05; **p<0.01; ***p<0.005 by Student's t test. B. For measurement of reactive oxygen species (ROS), cells were loaded with 30 µM of oxidant-sensitive DCF-DA dye after exposure to iron and fluorescent emission was measured at 540 nm. The experiment was performed in triplicates and repeated three times. The data are expressed as change in fluorescence/8×104 cells (± SEM). Iron uptake increased generation of ROS secretion in non-phagocytosing macrophages in all polarization states, most prominently in M1 polarization. Myelin-laden macrophages did not respond to iron exposure. *p<0.05; **p<0.005 by Student's t test. C. To visualize ROS generation, macrophages seeded on coverslips were incubated with 30 µM FeCl3 for 2 hrs and subsequently loaded with 30 µM of DCF-DA. Images were acquired at timed intervals (representative images at 10 min are shown). The results recapitulate ROS quantification in (B): ROS was increased in iron-laden MDMs [d–f] compared to naïve [a–c] or myelin-laden MDMs [g–i]. In contrast, labeling for the M2 polarization marker CD206, revealed decreased expression of CD206 in macrophages exposed to iron [dd–ff] and increased expression in myelin-laden macrophages [gg–ii].

Figure 5. FLAIR (3T) and GRE phase (7T) images of a patient with active relapsing-remitting MS.

FLAIR images show numerous white matter MS lesions of which 2 are magnified (inset, red arrows). Phase imaging at 7T phase/GRE reveals a hypointense ring corresponding with one lesion on FLAIR. The other lesion is not visible on 7T GRE (inset, arrows).

Figure 6. Prevalence of iron-positive lesions in active RR-MS and in SP-MS patients.

Patients were imaged with 3T (FLAIR) and 7T (GRE phase) and white matter MS lesions on FLAIR and on GRE phase images were quantified. The percentage of hypointense lesions (all hypointense lesions or lesions with hypointense rim only) that were also visible on corresponding FLAIR images was significantly higher in active RR-MS compared to SP-MS [A, B]. Similarly, the average numbers of phase-positive lesions visible on FLAIR (phase+/FLAIR+) and phase-positive lesions not visible on FLAIR (phase+/FLAIR−) was significantly higher in active RR vs. SP patients. In contrast, white matter MS lesions on FLAIR that were phase isointense (phase−/FLAIR+) did not differ significantly between patient groups [C]. (n = 8 patients/group; P values: (A) p = 0.009; (B) p = 0.0136. Data are expressed as means ± SEM).