|Home | About | Journals | Submit | Contact Us | Français|
Focal cerebral ischemia leads to an inflammatory reaction involving an overexpression of the peripheral benzodiazepine receptor (PBR)/18-kDa translocator protein (TSPO) in the cerebral monocytic lineage (microglia and monocyte) and in astrocytes. Imaging of PBR/TSPO by positron emission tomography (PET) using radiolabeled ligands can document inflammatory processes induced by cerebral ischemia. We performed in vivo PET imaging with [18F]DPA-714 to determine the time course of PBR/TSPO expression over several days after induction of cerebral ischemia in rats. In vivo PET imaging showed significant increase in DPA (N,N-diethyl-2-(2-(4-(2-fluoroethoxy)phenyl)-5,7-dimethylpyrazolo[1,5-a]pyrimidin-3-yl)acetamide) uptake on the injured side compared with that in the contralateral area on days 7, 11, 15, and 21 after ischemia; the maximal binding value was reached 11 days after ischemia. In vitro autoradiography confirmed these in vivo results. In vivo and in vitro [18F]DPA-714 binding was displaced from the lesion by PK11195 and DPA-714. Immunohistochemistry showed increased PBR/TSPO expression, peaking at day 11 in cells expressing microglia/macrophage antigens in the ischemic area. At later times, a centripetal migration of astrocytes toward the lesion was observed, promoting the formation of an astrocytic scar. These results show that [18F]DPA-714 provides accurate quantitative information of the time course of PBR/TSPO expression in experimental stroke.
Cerebral ischemia induces inflammatory processes triggered by reactions of resident glial cells and leukocyte infiltration. Reactive gliosis involves a set of stereotypic changes in the physiological state of microglia and astrocytes that is directly associated with the degree of brain damage (Kreutzberg, 1996). The activation of microglia includes changes in their morphology, migration toward the lesion site, proliferation, and acquisition of new functions in injured tissue, including the capacity to express and release a wide variety of proinflammatory molecules and to phagocytose dead cells. Conversely, reactive astrocytes form a rim surrounding the core of the infarction (Rojas et al, 2007), which may be responsible for protection of injured tissue (Maeda et al, 2007). The inflammatory reaction involves a dramatic increase in the expression of a mitochondrial transmembrane protein, the peripheral-type benzodiazepine receptor (PBR), whose upregulation is considered a hallmark of neuroinflammation (Chen and Guilarte, 2008). PBR is the original name for the binding site of isoquinoline PK11195, but a new name, translocator protein (18kDa) TSPO, has recently been introduced (Papadopoulos et al, 2006) to represent more accurately the complex functions of this protein. PBR/TSPO is an attractive target for imaging cerebral inflammation because of its very low level of expression in intact cerebral tissue (Chen et al, 2004). The targeting of PBR/TSPO with radiolabeled ligands has been reported in vitro in animal models of focal (Myers et al, 1991a, 1991b) and global (Stephenson et al, 1995) cerebral ischemia. The isoquinoline carboxamide, PK11195, the first non-benzodiazepine ligand specifically binding to PBR/TSPO, has been used widely for functional characterization of this protein and for identification of its cellular origin in brain tissue. [11C]PK11195 has been used as a radioligand in positron emission tomography (PET) studies to image active pathology in vivo in rodents (Rojas et al, 2007) and in human brain after cerebral ischemia (Gerhard et al, 2000; Pappata et al, 2000; Price et al, 2006), as well as neurodegenerative diseases (Banati et al, 2000; Cagnin et al, 2001). Nevertheless, [11C]PK11195 presents numerous limitations (Petit-Taboué et al, 1991), including high level of nonspecific binding, poor signal-to-noise ratio, and labeling with carbon-11 that limits its extensive clinical use (Chauveau et al, 2008). Among the alternative PET radioligands for PBR/TSPO, a new fluorine-18-labeled pyrazolo[1,5-a]pyrimidin-3-yl)acetamide, [18F]DPA-714, has recently demonstrated high affinity for PBR/TSPO in an ex vivo autoradiography of rat brain and a suitable biodistribution in baboons using PET imaging (James et al, 2008). Direct in vivo comparison of models of rats with acute neuroinflammation showed decreased nonspecific uptake and improved bioavailability of [18F]DPA-714 over [11C]PK11195, leading to higher binding potential in brain tissue (Chauveau et al, 2008). These promising results have led to preliminary clinical PET studies with [18F]DPA-714 for the detection of microglial activation in amyotrophic lateral sclerosis (Le Pogam et al, 2008).
Therefore, we reasoned that [18F]DPA-714 could document the time course of the neuroinflammatory reaction after stroke. We performed serial PET imaging with [18F]DPA-714 from 1 to 30 days after experimental middle-cerebral-artery (MCA) occlusion in rats. PET was conducted in parallel with autoradiography and immunohistochemistry of rat brains.
Adult male Sprague–Dawley rats (300g body weight; Charles River, Saint Germain sur L'arbresle, France) (n=106) were used. Animal studies were approved by the animal ethics committee and conducted in accordance with the Directives of the European Union on animal ethics and welfare. Transient focal ischemia was produced by a 2-h intraluminal occlusion of the MCA, followed by reperfusion, as described elsewhere (Justicia et al, 2006). Briefly, rats were anesthetized with 4% isoflurane in 100% O2. A 2.6-cm length of 4-0 monofilament nylon suture, which had been heat blunted at the tip, was introduced into the right external carotid artery up to the level where the MCA branches out. After occlusion, anesthesia was discontinued and rats were allowed to recover. After 2h, the animals were reanesthetized, the filament was removed, and the clip on the common carotid artery was released to allow reperfusion. Animals were studied at 1 (n=11), 4 (n=10), 7 (n=37), 11 (n=12), 15 (n=9), 21 (n=8), and 30 (n=10) days after reperfusion following the episode of ischemia. Nine non-operated (control) animals were used. Altogether, 57 animals were used to perform PET, 41 animals for in vitro autoradiographic studies and eight for immunohistochemical evaluation.
DPA-714 (N,N-diethyl-2-(2-(4-(2-fluoroethoxy)phenyl)-5,7-dimethylpyrazolo[1,5-a]pyrimidin-3-yl)acetamide) was labeled with fluorine-18 (half-life 109.8mins) at its 2-fluoroethyl moiety using a tosyloxy-for-fluorine nucleophilic aliphatic substitution according to slight modifications of procedures already reported (James et al, 2008). This simple one-step process has been automated on our Zymate-XP robotic system (Damont et al, 2008) and implemented on a commercially available GE TRACERLab FX-FN synthesizer. The process involves (a) reaction of K[18F]F-Kryptofix222 with the tosyloxy precursor (4.5 to 5.0mg, 8.2 to 9.1μmol) at 165°C for 5mins in dimethylsulfoxide (0.6mL), followed by (b) C-18 PrepSep cartridge prepurification, and finally (c) semi-preparative high-performance liquid chromatography purification on a Waters X-Terra RP18. Final formulation of [18F]DPA-714 as an intravenously injectable solution (physiological saline containing less than 10% of ethanol) was performed using a home-made SepPak Plus C18 device. Typically, 5.6 to 7.4GBq of [18F]DPA-714 (>95% chemically and radiochemically pure) was routinely obtained with specific radioactivities ranging from 37 to 111GBq/μmol within 85 to 90mins (high-performance liquid chromatography purification and SepPak-based formulation included), starting from a 37-GBq [18F]fluoride batch (overall non-decay-corrected and isolated radiochemical yield: 15 to 20%).
PET studies were performed for a total of 57 animals 1 (n=5), 4 (n=4), 7 (n=24), 11 (n=6), 15 (n=4), 21 (n=5), and 30 (n=6) days after induction of cerebral ischemia, as well as for non-operated rats used as controls (n=3). Of the 24 animals used at 7 days, eight were used to perform the PET time-course experiment, five for kinetics study, five for displacement using cold PK11195, and six for displacement using cold DPA-714. PET studies were not performed repeatedly over time for the same animal. Anesthesia was induced with 4% isoflurane and maintained by 2 to 2.5% of isoflurane in a mixture of 100% O2. During imaging the rat's head was placed in a home-made stereotaxic' frame compatible with PET acquisition and rats were kept normothermic using a heating blanket (Homeothermic Blanket Control Unit; Harvard Apparatus Limited, Edenbridge, Kent, UK). The tail vein was catheterized with a 24-gauge catheter for intravenous administration of all compounds. PET imaging was performed on a Concorde Focus 220 camera (spatial resolution 1.35-mm full width at half maximum) dedicated to small animal imaging. [18F]DPA-714 (74.1MBq) was injected concomitantly with the start of PET acquisition. For displacement experiments, unlabeled compounds (PK11195 and DPA-714, 1mg/kg) were injected 20mins after the injection of radiotracers. PET data were acquired over 60mins after radiotracer injection in the case of kinetics and displacement studies performed 7 days after ischemia, and over 30mins at all other postischemia time points. The attenuation correction factors were measured using an external 68Germeniun point source.
Finally, the emission sinograms were normalized, corrected for attenuation and radioactivity decay, and reconstructed using FORE and OSEM 2D (16 subsets and four iterations). PET images with [18F] DPA-714 were reconstructed with 17 or 24 dynamic frames.
PET images were manually co-registered to the anatomical data of an MRI template for rat brain (Schweinhardt et al, 2003) with the aim of localizing anatomically the PET signal using Anatomist software (http://brainvisa.info).
Regions of interest (ROIs) were manually defined for each rat on the region of increased binding in the ipsilateral hemisphere. A spherical ROI containing the major part of the territory irrigated by the MCA was defined in the homologous contralateral hemisphere. The mean value and the standard deviation of each ROI in the PET images were measured using the Anatomist software. Values were expressed as percentage of injected dose per cubic centimeter (ID/cc) for each ROI.
A group of 41 rats were used at 1 (n=5), 4 (n=5), 7 (n=12), 11 (n=5), 15 (n=4), 21 (n=2), and 30 (n=3) days after ischemia, together with control animals (n=5). The animals were terminally anesthetized and killed by decapitation. The brain was removed from the skull, immediately frozen in isopentane in liquid nitrogen, and sliced in 20-μm coronal sections, which were used immediately for autoradiographic studies of [18F]DPA-714 (74.1MBq). Using adjacent sections, specific binding at 7 days after ischemia was assessed with an excess (20μmol) of unlabeled PK11195 and DPA-714. Sections were incubated for 20mins in Tris Buffer (TRIZMA pre-set Crystals (Sigma, Saint Quentin Favallier, France) adjusted at pH 7.4 at 4°C, 50mM, with NaCl 120mM) and then rinsed for 2mins with cold buffer, followed by a quick wash in cold distilled water. Sections were then placed in direct contact with a Phosphor-Imager screen and exposed overnight. Autoradiograms were analyzed using the ImageQuant software. ROIs of interest were manually defined for each rat on the region of increased binding in the ipsilateral hemisphere, and a spherical ROI containing the major part of the territory irrigated by the MCA was defined in the homologous contralateral hemisphere. The mean value and the standard deviation of each ROI in autoradiograms were expressed as counts per cubic centimeter in the time-course binding study, and as the ratio of counts per cubic centimeter of the ischemic area in relation to the contralateral hemisphere in displacement studies.
The simplified reference-tissue model (SRTM) (Lammertsma and Hume, 1996) from the PMOD software package (version 2.5; PMOD Technologies Ltd, Zürich, Switzerland) was used to assess binding potential (BP) in the ipsilateral ROI. This model relies on a two-tissue reversible compartment for the target region (ipsilateral ROI) and a single-tissue compartment for the reference region (contralateral ROI). Three parameters were estimated for each kinetic: R1 (K1/K′1), which represents the ratio of tracer delivery; k2, which is the clearance from the target tissue back to the vascular compartment; and BP (k3/k4), which is the binding potential of the tracer to the tissue. PET data modeling was performed at days 1 (n=3), 4 (n=4), 7 (n=7), 11 (n=6), 15 (n=4), 21 (n=5), and 30 (n=4) after ischemia using data from the PET time-course study.
Immunohistochemical staining was performed for each group of rats (n=8) at different time points after ischemia (one animal per time point and one control). Animals were terminally anesthetized, killed by decapitation, and the brain was removed from the skull. The brains were frozen and 5-μm-thick frozen sections were generated in a cryostat. One set of rat brain sections was fixed in paraformaldehyde (PFA) 4% for 15mins. PFA blockage was performed with NH4Cl (50mmol/L) in phosphate-buffered saline (PBS) for 5mins and tissue permeabilization was performed with methanol-acetone 1:1 (−20°C) and Triton 0.1% in PBS for 5mins. Sections were washed with PBS after every step. The sections were saturated with a solution of bovine serum albumin (5%)/Tween (0.5%) in PBS for 15mins at room temperature to block nonspecific binding. Sections were incubated for 1h at room temperature with primary antibodies in bovine serum albumin (5%)/Tween (0.5%) in PBS. Sections were stained for glial fibrillary acidic protein (GFAP) with rabbit anti-rat GFAP (1:500; DakoCytomation, Trappes, France), for CD11b (Ox42) with mouse anti-rat CD11b (1:300; Serotec, Raleigh, NC, USA), and finally for peripheral benzodiazepine receptor with a rabbit anti-rat PBR/TSPO (NP155, 1:1000) (Ji et al, 2008). The sections were then washed (3 × 10mins) in PBS and incubated for 1h at room temperature with secondary antibodies Alexa Fluor 488 and 750nm goat anti-rabbit IgG and Alexa Fluor 594nm goat anti-rat IgG (1/1000; Molecular Probes, Invitrogen, Les Ulis, France) in bovine serum albumin 5%/Tween 0.5% in PBS, and then washed again (3 × 10mins) in PBS. Sections were mounted with a 4,6-diamidino-2-phenylindole Prolong Antifade kit (Molecular Probes, Invitrogen).
The number of GFAP/TSPO and CD11b/TSPO immunopositive cells within the ischemic area was assessed by counting the number of stained cells at 1, 4, 7, 11, 15, 21, and 30 days after ischemia. Cells were counted in five different fields totaling an area of 7mm2 at × 63 magnification. Images of areas showing the highest staining density were acquired and cells were counted manually.
Differences at various time points after ischemia versus control were analyzed using one-way analysis of variance, and Bonferroni's multiple comparison tests were used for post hoc analysis. Two-group comparison was evaluated by an unpaired t-test.
An increase in [18F]DPA-714 binding was detected with PET at days 7, 11, 15, and 21 after ischemia in the stroke area (ipsilateral) with respect to the control area (Figures 1A to 1G). The time course of the [18F]DPA-714 PET signal was measured in the ipsilateral and in the contralateral hemisphere (see section Methods) at 1, 4, 7, 11, 15, 21, and 30 days after induction of cerebral ischemia. Increased [18F]DPA-714 binding was detected in ipsilateral areas of ischemic brains at days 7, 11, 15 (P<0.01), and 21 (P<0.05, with respect to control animals). A significant increase in [18F]DPA-714 uptake was found at days 7 (P<0.05) and 11 (P<0.001 versus day 4 after ischemia). The highest binding value was reached at 11 days after ischemia, compared with 7 and 15 days (P<0.05). After day 15, the PET signal decreased at 21 days with respect to 11 days (P<0.01), and decreased further at day 30 after ischemia in relation to 11 days (P<0.001) and 15 days (P<0.05) (Figure 1H). The [18F]DPA-714 PET signal in the contralateral hemisphere showed significant increase versus that in the controls at 11 days after ischemia (P<0.05; Figure 1I).
The time–activity curve generated at day 7 after ischemia showed that [18F]DPA-714 uptake in the lesion reached a peak value at 30mins and remained at this value until 60mins after the bolus injection. In the contralateral hemisphere, the uptake showed a peak of radioactivity followed by fast washout (Figure 2A).
Displacement studies were performed by injecting an excess (1mg/kg) of either PK11195 or DPA-714 20mins after the tracer injection. PET images co-registered with an MRI rat brain atlas localized the lesion area and showed the reduction in [18F]DPA-714 uptake after displacements by DPA-714 and PK11195 (Figure 3). Five to 10mins after the cold-compound injection, the radioactivity concentration in the stroke area decreased to the concentration levels of the contralateral area (Figures 2B and 2C). Accordingly, the ratios of ipsilateral-to-contralateral brain regions showed a decrease in the signal after displacement by DPA-714 and PK11195 at 7 days after ischemia (P<0.01). There was a slightly higher decrease in the signal after displacement with DPA-714, as compared with PK11195, but the difference was not statistically significant (Figure 2D). The injection of DPA-714 (Figure 2C) induced faster displacement of [18F]DPA-714 binding than did the injection of PK11195 (Figure 2B).
In a first series of experiments, the time course of [18F]DPA-714 binding tissue was performed with brain sections of rats killed 1, 4, 7, 11, 15, 21, and 30 days after the induction of cerebral ischemia, to verify the results obtained in vivo with PET. [18F]DPA-714 binding was significantly increased in the ipsilateral area at days 4 (P<0.01), 7, 11, and 15 (P<0.001), and at day 21 (P<0.05 with respect to the control). Values at days 7 (P<0.05) and 11 (P<0.001) were also significantly higher than those at 4 days after ischemia. The highest binding value was reached 11 days after ischemia, as compared with that at 7 and 15 days (P<0.05). After day 15, the binding decreased at 21 days with respect to binding at 11 (P<0.001) days, and decreased further at day 30 after ischemia with respect to 11 (P<0.001) and 15 days (P<0.05; Figure 4A). With respect to the time interval after induction of the ischemic lesion in the brain, autoradiographic measurements of [18F]DPA-714 binding in the contralateral hemisphere (data not shown) showed a time course similar to that of in vivo [18F]DPA-714 uptake measured with PET.
The second set of experiments was performed to confirm the displacement studies using PET. In competition binding studies, brain sections of rats collected 7 days after ischemia were co-incubated with [18F]DPA-714 and a 1000 to 10,000-fold excess of PK11195 (20μmol) or DPA-714 (20μmol). Ratios of lesioned-to-contralateral brain hemisphere showed a decrease in the signal after both displacements (P<0.001). The results suggested a higher displacement of the signal by DPA-714 than by PK11195, although, for in vivo PET results (Figure 4B), this difference was not statistically significant.
PET data were modeled using a SRTM (Lammertsma and Hume, 1996 and Figure 5). The BP of [18F]DPA-714 in the lesioned area increased at day 7 (P<0.01) and 11 (P<0.001 with respect to day 1; Figure 5A). BP values were also significantly higher at 7 days than at days 4, 15 (P<0.01), 21 (P<0.05), and 30 (P<0.01). The ratio of the transport rate from blood to brain tissue in lesion versus the reference region (R1) was close to unity at day 1, but increased at days 7 (P<0.01) and 11 (P<0.05) with respect to day 1 (Figure 5B). The clearance (k2) of [18F]DPA-714 from the lesioned area showed a decrease at days 4 (P<0.05), 7 (P<0.001), 11 (P<0.05), 15, 21, and 30 (P<0.001 with respect to day 1; Figure 5C).
Immunofluorescence staining at different times after cerebral ischemia showed PBR/TSPO expression in two glial subpopulations, microglia and astrocytes. At day 4, cells with the morphology of amoeboid reactive microglia/macrophages showed intense CD11b/TSPO immunoreactivity in the lesion (Figures 6A to 6M). At 7 days, GFAP-positive astrocytes showing expression for PBR/TSPO appeared in the vicinity of the lesion, forming a thin astrocytic rim (data not shown). From day 11 onwards, immunofluorescence staining showed astrocytic migration from the periphery toward the lesion, itself mainly occupied by a massive accumulation of reactive microglia accompanied by high expression of PBR/TSPO (Figures 6B to 6N). At 30 days, an ischemic area encircled by numerous GFAP/TSPO-positive astrocytes, delineating a consolidated astrocytic scar, was observed. Astrocytes showed close contact with numerous CD11b/TSPO-positive microglia/macrophages (Figures 6C to 6O). In addition, immunolabeling showed clear decrease in PBR/TSPO expression 11 days after ischemia (Figures 6I and 6H). In the contralateral area, a dispersed and very modest increase in PBR/TSPO expression was observed only at 11 days after ischemia (data not shown).
Immunohistochemistry for microglia and astrocytes illustrated the localization of glial subtypes in injured tissue. A time course of TSPO+/CD11b+ microglia/macrophage and TSPO+/GFAP+ was performed in the ischemic lesion at 1, 4, 7, 11, 15, 21, and 30 days. An increase in TSPO+/CD11b+ cells was detected at days 7, 11, 15, 21 (P<0.001 with respect to the control), and 1. Also, an increase in TSPO+/CD11b+ cells was found at day 30 (P<0.01) in relation to the control (Figure 7A). An increase in TSPO+/CD11b+ was also observed at days 7, 11, and 15 (P<0.001 versus 4 days after ischemia). The highest number of TSPO+/CD11b+ cells was reached at 11 days after ischemia compared with 21 (P<0.05) and 30 days (P<0.01). At day 30, the number of TSPO+/CD11b+ cells also decreased in relation to 7 and 15 days (P<0.05). The number of TSPO+/GFAP+ cells was increased in the lesion at day 21 in relation to the control, and at days 1 (P<0.001), 4 (P<0.01), and 7, and 11 (P<0.05). The highest level of TSPO+/GFAP+ cells was observed at day 30 (P<0.001 with respect to the control, days 1, 4, and 7; P<0.01 with respect to day 11; Figure 7B).
This PET study evaluates PET imaging with [18F]DPA-714, a new PBR/TSPO radioligand, in experimental cerebral ischemia. Previously, [18F]DPA-714 uptake was evaluated ex vivo in striatal lesions in rat brains, and in vivo, in healthy baboons (James et al, 2008), in a rat model of herpes encephalitis (Doorduin et al, 2009), and in acute neuroinflammation in rats (Chauveau et al, 2009). Under physiological conditions, brain parenchyma shows low expression of PBR/TSPO, but it is dramatically upregulated under neuropathological conditions, including neurodegenerative diseases, brain trauma, and cerebral ischemia. Overexpression of PBR/TSPO has been related to the activation of glial cells after brain injury (Ji et al, 2008). PBR/TSPO ligands have been involved in the modulation of chemotaxis, phagocytosis and steroidogenesis. Therefore, the microglial response to injury, such as migration, proliferation, and phagocytosis, may be related to PBR/TSPO overexpression. Moreover, under certain circumstances, PBR/TSPO overexpression can promote mitochondrial permeability (Chelli et al, 2001) and apoptotic processes (Castedo et al, 2002). In contrast, activation of PBR/TSPO in astrocytes after injury has been related to synthesis of pregnenolone and progesterone, neurosteroids that perform neurotrophic and neuroprotective activities (Le Goascogne et al, 2000; Schumacher et al, 2000). In addition, PBR/TSPO participates in cholesterol and protein transport, porphyrin and heme biosynthesis, immunomodulation, and cellular respiration (Papadopoulos et al, 2006). Thereby, PBR/TSPO is considered an attractive and sensitive marker for the quantification and visualization of neuropathological changes induced during cerebral ischemia. In this study, a 2-h occlusion model of transient cerebral ischemia induced a time-dependent increase in [18F]DPA-714 PET signal in the lesion as compared with that in the contralateral side. In vivo PET imaging and in vitro autoradiographic studies confirmed increase in [18F]DPA-714 binding at 4 and 7 days after cerebral ischemia, reaching a maximal value at 11 days, followed by slow return to normal values at 30 days. An increased binding of [11C]PK11195 during the first week after cerebral ischemia in rodents has been reported in autoradiographic (Myers et al, 1991a, 1991b) and PET studies (Rojas et al, 2007). Similarly, clinical PET studies with [11C]PK11195 showed increased binding 1 week after stroke, which extended over time (Gerhard et al, 2005; Price et al, 2006). We also observed increased [18F]DPA-714 binding in the contralateral side at 11 days after cerebral ischemia, because of the further increase in binding observed in the lesion area, probably through different mechanisms related to the expansion of infarction, such as spreading depression (Jander et al, 2003). Contralateral binding was completely displaced using cold PK11195 and cold DPA-714 7 days after ischemia, strongly supporting the view that this modest and transient increase in [18F]DPA-714 binding is PBR/TSPO-specific. In PET studies, 7 days after ischemia, the time–activity curve for [18F]DPA-714 showed increased uptake in the ipsilateral side, which peaked 30mins after injection and was maintained at a plateau level during the following half hour. This result supports the use of PET scans of 30mins (0 to 30mins) duration for comparison of [18F]DPA-714 uptake at different time points after stroke. In the contralateral area, [18F]DPA-714 showed lower uptake because of low presence of PBR/TSPO in intact tissue, confirming a previous report that [18F]DPA-714 shows low nonspecific binding in intact tissue (Chauveau et al, 2009). [18F]DPA-714 binding was rapidly displaced from the ipsilateral area by an excess of the corresponding unlabeled compound, where, within minutes, its uptake was close to that in the non-injured area. The target-to-background ratio after displacement by DPA-714 and PK11195 was 1.5 and 1.7, respectively, showing incomplete displacement from the lesion. Accordingly, in vitro autoradiography confirmed the trend of a higher displacement capacity by DPA-714 than by PK11195, as observed by PET imaging. The better displacement showed by DPA-714 in relation to PK11195 may suggest a higher specificity of DPA-714 toward PBR/TSPO.
Modeling with SRTM showed that tracer delivery ratios (R1) were higher than 1 as early as day 4 after cerebral ischemia and onwards, indicating facilitated entry of the tracer in the lesion or in the ipsilateral area in relation to the contralateral area. These results correlate with the widespread breakdown of the blood–brain barrier at 48h after MCA occlusion (Belayev et al, 1996). BP values on different days after ischemia correlated with the PET binding time course. However, the BP value peaked at 7 days after ischemia, whereas [18F]DPA-714 concentrations peaked at 11 days. The earlier time at which BP peaked after ischemia (7 days) with respect to [18F]DPA-714 concentrations (11 days) could be explained either by (i) the larger dispersion of values observed at 7 days or (ii) by the increase in [18F] DPA-714 binding in the contralateral side (reference tissue) evidenced at 11 days, because of the spreading of inflammation-activating factors away from the ischemic zone, or (iii) both. In this study, modeling of PET data with SRTM suggested that, between 7 and 11 days after ischemia, there may have been increased clearance from the target tissue to the vascular compartment (k2 value), coincident with a decrease in the BP value. However, this may also reflect the limitations of the SRTM model and further studies using alternative compartmental modelization of data than SRTM may be necessary to better document the time course of PBR/TSPO while using [18F]DPA-714.
The inflammatory reactions after ischemia were characterized by immunohistochemistry at the same time points at which PET and autoradiography studies were conducted. The results showed an overall increase in PBR/TSPO expression in the ischemic area after cerebral ischemia and indicated that this increase was heterogeneous. The number of CD11b-reactive cells (i.e., of monocytic lineage) increased progressively up to day 11 and decreased in the following weeks, until day 30. This observation is consistent with reports of increased microglial populations at 7 days after ethanol-induced neuronal insults, followed by a later decrease at 30 days (Maeda et al, 2007). The expression of PBR/TSPO in astrocytes was observed in a rim surrounding the epicenter of the lesion at 4 and 7 days after cerebral ischemia, in agreement with the results observed by Rojas et al (2007) in a model of 1-h MCA occlusion. After 11 days, astrocytes started to show centripetal migration toward the core of the inflammatory lesion, without overlapping with the expression of PBR/TSPO in microglia/macrophages. The migration of astrocytes over time toward the lesion led to the formation of an astrocytic scar, which possibly acts as a retaining wall preventing spatial propagation of the damage caused by activated microglia. In our results, we observed increased PBR/TSPO activity in the astrocytes forming the scar, which could be related to their migration capacity and supports the possible beneficial role of PBR/TSPO in astrocytes, as suggested by Ji et al (2008).
Interestingly, the time course of [18F] DPA-714 binding that we observed using PET and autoradiography matched with the temporal profile of microglia/macrophage activation, but not with that of astrocytes expressing PBR/TSPO. These findings could be explained by (i) a higher expression of PBR/TSPO in microglia/macrophages than in astrocytes, even at late time points after ischemia or by (ii) the presence of different subtypes of PBR/TSPO in different glial cells, with different affinities for [18F]DPA-714.
In summary, we show here that, after acute ischemia, [18F]DPA-714 uptake is mainly associated with the expression of PBR/TSPO observed in microglia/macrophages over time. In addition, the population of astrocytes expressing PBR/TSPO increases after reperfusion of the ischemic lesion. [18F] DPA-714 provides accurate informative information of the expression and distribution of PBR/TSPO activity after cerebral ischemia. Nevertheless, a more precise localization of the increase in [18F]DPA-714 binding, PBR/TSPO expression, and inflammatory correlates, with respect to infarcted and peri-infarcted regions, remains to be further studied.
We thank Ms K Siquier, Mr B Jego, and Ms A Blossier for technical assistance, and Dr M Higuchi the gift of antibody NP155. This study was funded by the EC-FP6 network EMIL (LSHC-CT-2004-503569) and EC-FP6 network DiMI (LSHB-CT-2005-5121146).
Conflict of interest
The authors declare no conflict of interest.