Skip to main content
NCBI home page
Journal of Cerebral Blood Flow & Metabolism logoLink to Journal of Cerebral Blood Flow & Metabolism
. 2011 Apr 27;31(8):1807–1816. doi: 10.1038/jcbfm.2011.55

Quantitation of translocator protein binding in human brain with the novel radioligand [18F]-FEPPA and positron emission tomography

Pablo M Rusjan 1,*, Alan A Wilson 1,2, Peter M Bloomfield 1, Irina Vitcu 1, Jeffrey H Meyer 1,2, Sylvain Houle 1,2, Romina Mizrahi 1,2
PMCID: PMC3170950  PMID: 21522163

Abstract

This article describes the kinetic modeling of [18F]-FEPPA binding to translocator protein 18 kDa in the human brain using high-resolution research tomograph (HRRT) positron emission tomography. Positron emission tomography scans were performed in 12 healthy volunteers for 180 minutes. A two-tissue compartment model (2-CM) provided, with no exception, better fits to the data than a one-tissue model. Estimates of total distribution volume (VT), specific distribution volume (VS), and binding potential (BPND) demonstrated very good identifiability (based on coefficient of variation (COV)) for all the regions of interest (ROIs) in the gray matter (COV VT<7%, COV VS<8%, COV BPND<11%). Reduction of the length of the scan to 2 hours is feasible as VS and VT showed only a small bias (6% and 7.5%, respectively). Monte Carlo simulations showed that, even under conditions of a 500% increase in specific binding, the identifiability of VT and VS was still very good with COV<10%, across high-uptake ROIs. The excellent identifiability of VT values obtained from an unconstrained 2-CM with data from a 2-hour scan support the use of VT as an appropriate and feasible outcome measure for [18F]-FEPPA.

Keywords: inflammation, kinetic modeling, microglia, mitochondria, positron emission tomography

Introduction

Microglia are key players in the immune surveillance system in the central nervous system where they are the resident macrophages (Gehrmann et al, 1995) and are the first responders to various types of brain injury (Kreutzberg, 1996). Microglia express a protein in their mitochondria called the translocator protein 18 kDa (TSPO) (Braestrup et al, 1977). The TSPO previously known as the peripheral benzodiazepine receptor (PBR) is located in the outer mitochondrial membrane (Bernassau et al, 1993; McEnery et al, 1992) and is part of a larger complex, which includes the adenine nucleotide carrier and the voltage-dependent anion channel (Culty et al, 1999; Papadopoulos et al, 1997). In response to neuroinflammation, TSPO is overexpressed compared with normal tissue. This has been confirmed in a large number of studies over several years, which have shown that levels of TSPO and/or microglia are greatly increased in examples of inflammation (as reviewed by Chen and Guilarte, 2008).

The first and most widely used radiotracer in positron emission tomography (PET) imaging of TSPO is [11C]-PK11195 (Cagnin et al, 2001; Camsonne et al, 1984; Charbonneau et al, 1986; Debruyne et al, 2003; Gerhard et al, 2000, 2003; Pappata et al, 1991). However, this radiotracer has recognized limitations including high nonspecific binding, low brain penetration, high plasma protein binding, and a difficult synthesis. The deficiencies of [11C]-PK11195 coupled with the recognized importance of TSPO imaging have fueled considerable efforts to develop radiotracers with greater sensitivity to detect TSPO binding (James et al, 2006; Okubo et al, 2004). While [11C]-PBR28 (Fujita et al, 2008) and [11C]-DPA713 (Boutin et al, 2007) circumvent many of the deficiencies of [11C]-PK11195, both are radiolabeled with the short-lived carbon-11 (t1/2=20.39 minutes), which make them unsuitable for wide-spread dissemination. The advantages of fluorine-18-labeled radiotracer are several; improved targetry means that [18F]-fluoride can now be produced by low/medium energy cyclotrons in large quantities, the imaging quality of this radionuclide is superior to carbon-11, and most importantly the longer half-life (t1/2=109.8 minutes), allows shipment to imaging sites distant from the site of production, which is particularly useful for clinical applications. Only two [18F]-labeled TSPO radiotracers have been evaluated so far in humans. [18F]-FEDAA1106 displays good brain penetration, but this compound is exceedingly lipophilic and its binding kinetics are very slow (Fujimura et al, 2006). Very recently, [18F]-PBR06 has been studied in human (Fujimura et al, 2009). [18F]-PBR06 has many favorable properties, including appropriate kinetics, good brain penetration, and ease of preparation. Unfortunately, initial reports suggest that it produces a brain-penetrant radiolabeled metabolite, which confounds quantification of TSPO binding (Fujimura et al, 2009).

We recently reported the radiosynthesis and initial evaluation of an F-18 radiolabeled analog of PBR28 (Wilson et al, 2008) and compare its in vitro, physio-chemical, and in vivo central nervous system distribution properties, including metabolism in rats, to PBR28 (Wilson et al, 2008). FEPPA (Ki≈0.07 nmol/L) was threefold more potent than PBR28 (Ki≈0.22nmol/L) and an order of magnitude more potent than either DPA713 (Ki≈0.87 nmol/L) or the prototypical PK11195 (Ki≈1.29 nmol/L) (Wilson et al, 2008). The lipophilicity of FEPPA was measured at physiological pH and found to be 2.99 (Wilson et al, 2008), suitable for penetration of the blood–brain barrier. In pig, [18F]-FEPPA showed good and rapid brain penetration with an appropriate regional distribution for binding to TSPO. Highest uptake was seen in the thalamus with lower amounts in the cerebellum and frontal cortex. Pretreatment with the prototypical TSPO ligand PK11195 (2 mg/kg) resulted in a significantly lower and homogeneous binding pattern, demonstrating that a large proportion of the brain uptake of [18F]-FEPPA is mediated by binding to TSPO (Bennacef et al, 2008). In summary, [18F]-FEPPA displays highly favorable properties as a radiotracer for PET imaging of TSPO/neuroinflammation in human as it has high affinity for TSPO, an appropriate metabolic profile, with high brain penetration and good pharmacokinetics in both pig and rat. Finally, [18F]-FEPPA is demonstrably sensitive to neuroinflammation in a rat model (Hatano et al, 2010).

In light of these promising results, we evaluated the ability of [18F]-FEPPA to quantify TSPOs in the human brain. We sought to determine whether brain uptake was better quantified with a model having a single compartment (i.e., free, specifically bound, and nonspecifically bound radiotracer instantaneously in equilibrium) or two compartments (i.e., in the first compartment free and nonspecifically bound radiotracer instantaneously in equilibrium and in the second specifically bound radiotracer). We also investigated (1) how well the total distribution volume (VT), the distribution volume of the specifically bound radiotracer (VS), and the binding potential (BPND) were identified, (2) whether stable values of VT and BPND were attained during the scanning session, and (3) how much of an increase of TSPO binding can reliably be measured during a 2-hour scanning session.

Materials and methods

Radiosynthesis of [18F]-FEPPA

Details of [18F]-FEPPA radiosynthesis have been described previously (Wilson et al, 2008). Briefly, [18F]-fluoride is dried, then reacted with the tosylate precursor in acetonitrile for 10 minutes at 90°C. The product is purified by high-performance liquid chromatography and formulated in buffered saline containing 5% to 10% ethanol, then cold sterilized by passing through a 0.22-μ filter. The final formulation is sterile, pyrogen free, with a pH of 5 to 8. Radiochemical purity is >96% with specific activities of above 1,000 mCi/μmol at the end of synthesis. The formulation is stable, apart from radioactive decay, for at least 6 hours. The whole radiosynthetic procedure is performed in a General Electric Medical systems FXN (Milwaukee, WI, USA) sealed module.

Human Subjects

Twelve healthy volunteers (four men and eight women; age 24 to 72 years) participated in this study. All subjects were free of current medical and psychiatric illness based on history, physical examination, electrocardiogram, urinalysis (including drug screening), and blood tests (complete blood count and serum chemistry). All subjects provided written informed consent after all procedures were fully explained.

Positron Emission Tomography Protocol

The PET scanning was performed using a 3D high-resolution research tomograph (HRRT) brain tomograph (CPS/Siemens, Knoxville, TN, USA), which measures radioactivity in 207 slices with an interslice distance of 1.22 mm. The detectors of the HRRT are an LSO/LYSO phoswich detector, with each crystal element measuring 2 × 2 × 10 mm3. A transmission scan, measured using a single photon point source, 137Cs (t1/2=30.2 years, Eγ=662 keV) was acquired immediately before the acquisition of the emission scan. This transmission scan was subsequently used to correct the emission data for the attenuation of the emission photons through the head and support.

Each subject was scanned for 180 minutes following the injection of [18F]-FEPPA, and the images were reconstructed into 45 time frames: The first frame was of variable length being dependent on the time between the start of acquisition and the arrival of [18F]-FEPPA in the tomograph field of view (FOV). The subsequent frames were defined as 5 × 30 seconds, 14 × 5 seconds, 2 × 60 seconds, 1 × 90 seconds, 1 × 120 seconds, 1 × 210 seconds, and 33 × 300 seconds.

The emission list mode data were rebinned into a series of 3D sinograms. The 3D sinograms were gap filled, scatter corrected, and Fourier rebinned into 2D sinograms The images were reconstructed from the 2D sinograms using a 2D filtered-back projection algorithm, with a HANN filter at Nyquist cutoff frequency. The reconstructed image has 256 × 256 × 207 cubic voxels measuring 1.22 × 1.22 × 1.22 mm3 and the resulting reconstructed resolution is close to isotropic 4.4 mm, full width at half maximum in plane and 4.5 mm full width at half maximum axially, averaged over measurements from the center of the transaxial FOV to 10 cm off-center in 1.0 cm increments.

A custom-fitted thermoplastic mask was made for each subject and used with a head fixation system during PET measurements. In addition, head movement was corrected after the scan by realigning all images from each subject using Statistical Parametric Mapping (version 5 (SPM5); Wellcome Department of Cognitive Neurology).

Measurement of [18F]-FEPPA in Plasma

Arterial sampling was taken continuously at a rate 2.5 mL/min for the first 22.5 minutes. The continuous early arterial blood radioactivity levels were counted using an automatic blood sampling system (Model # PBS-101 from Veenstra Instruments, Joure, The Netherlands). In addition, 4 to 10 mL manual samples were taken at 2.5, 7, 12, 15, 20, 30, 45, 60, 90, 130, and 180 minutes. An aliquot of each blood sample was taken to measure radioactivity concentration in total blood. The remaining blood was centrifuged (1,500 g, 5 minutes) and a plasma aliquot counted together with the total blood sample using a Packard Cobra II γ counter crosscalibrated with the PET system. The blood-to-plasma ratios were determined from the manual samples to correct the blood radioactivity time–activity curve (TAC) measured by automatic sampling and to generate the plasma radioactivity curve. A biexponential function was used to fit the blood-to-plasma ratios. The remaining volume of each manual plasma sample was used to determine parent radioligand and its metabolites in plasma. A Hill function was used to fit the percentage of unmetabolized tracer. A metabolite corrected plasma curve was generated by the product of the dispersion corrected blood curve with the two curves (blood-to-plasma ratio and percentage of parent radiotracer), which was then used as the input function for the kinetic analysis.

Magnetic Resonance Image and Regions of Interest Delineation

For the anatomical delineation of regions of interest (ROIs), a brain magnetic resonance image was acquired for each subject. 2D axial proton density images were acquired with a General Electric (Milwaukee, WI, USA) Signa 1.5 T magnetic resonance image scanner (slice thickness=2 mm, repetition time>5,300 milliseconds, echo time=13 milliseconds, flip angle=90°, number of excitations (NEX)=2, acquisition matrix=256 × 256, and FOV=22 cm). Regions of interest for the cerebellar cortex (hereafter referred to as the cerebellum), caudate, putamen, frontal cortex, temporal cortex, occipital cortex, insular cortex, anterior cingulate cortex, and thalamus were automatically generated based in those proton density-magnetic resonance images using in-house software, ROMI (Rusjan et al, 2006). ROMI utilizes computer vision techniques based on the probabilities of gray matter to fit a standard template of ROIs to an individual high-resolution magnetic resonance image scan. A ROI for white matter is generated using a previously described algorithm (Bencherif et al, 2004). The individual magnetic resonance images are then registered to a time average of the dynamical PET image so that the individual refined ROI template is transformed to the PET image space to allow the TAC generation from each ROI. Coregistration was performed using SPM2 (Welcome Department of Cognitive Neurology, London), which optimizes a measure of normalized mutual information (Studholme et al, 1999).

Kinetic Analysis

Following the definitions proposed in a consensus nomenclature for reversibly binding radioligands (Innis and Carson, 2007), TAC data were analyzed with one-tissue compartment model (1-CM) to estimate rate constants K1 and k1, two-tissue compartment model (2-CM) to estimate K1, k2, k3, and k4 and Logan graphical analysis to estimate VT.

VT is equal to the ratio at equilibrium of the concentration of radioligand in tissue to that in plasma. VT includes the concentrations of all radioligand in tissue (i.e., specific binding and nondisplaceable uptake (nonspecifically bound and free radioligand in tissue)). The value of VT can be estimated from the rate constants, for 1-CM as VT1-CM=k1/k2 and for 2-CM as VT2-CM=k1/k2 (1+k3/k4). In addition, for the 2-CM it is possible to directly estimate the distribution volume of the specific compartment VS (i.e., the ratio at equilibrium of the specifically bound radioligand to that of total parent radioligand in plasma) as VS=K1/k2 k3/k4 and BPND (i.e., a ratio at equilibrium of specifically bound radioligand to that of nondisplaceable radioligand in tissue (Mintun et al, 1984)) as BPND=k3/k4.

Nonlinear Least-Square Fitting

Kinetic analyses were performed using PMOD 3.1 modeling software (PMOD Technologies Ltd., Zurich, Switzerland) (Burger and Buck, 1997). Rate constants were estimated with the weighted least-squares method and the Marquardt optimizer. The nonlinear fitting for the 2-CM used as independent) variables K1, K1/k2, k3, and k4 for each ROI. The percent coefficient of variation (%COV=100% × s.e./mean) was used to measure the identifiability of the kinetic variables. The standard error (s.e.) was estimated from the diagonal of the covariance matrix of nonlinear least-squares fitting. A smaller percentage indicates better identifiability.

Brain data for each frame were weighted relative to other frames based on the trues (Ti) in the FOV during the frame i in the following way (Yaqub et al, 2006):

graphic file with name jcbfm201155e1.jpg

where λ=ln (2)/109.8 minutes is the decay constant of 18F, and tis and tie is the frame start and end time. Each model configuration was implemented to account for the contribution of activity from the cerebral blood volume assuming that cerebral blood volume was 5% of the brain volume (Phelps et al, 1979) for gray matter and 2.7% for white matter (Leenders et al, 1990). The whole blood–activity curve (Cb) was calculated correcting by delay and dispersion the curve measured with the automatic blood sampling system (Cm). The dispersion effect was modeled as Cm(t)=(1/τ) e−t/τ⊗Cb(t) with the time constant of dispersion τ=16 seconds calculated from previous experiments with our hardware. The delay, δ, between the activity in the FOV, described by the head curve (Hc=prompts–randoms) in a second by second basis (Iida et al, 1986; Meyer, 1989) and the plasma input function (Cp) was estimated by fitting the first 50 seconds of the Hc to an irreversible 1-CM with Cb as input function and including a term for the cerebral blood volume:

graphic file with name jcbfm201155e2.jpg

Simulations

Monte Carlo simulations were performed to assess the loss of identifiability in VT, VS, and BPND when k3 increases. The thalamus was chosen as the ROI to perform this simulation, as it presents the highest binding (i.e., least reversible curve) with an important level of noise, which would more likely mimic an unfavorable situation. The scanning length for each simulation was 2 hours. Noise for the frame i at time ti was modeled with a Gaussian distribution with standard deviation (s.d.i) according to Logan et al (2001):

graphic file with name jcbfm201155e3.jpg

where Ci is the noise-free simulated radioactivity and SF is the scale factor that controls the level of noise. Setting SF=1.25, the mean percent noise contained in the noisy data was calculated as the ratio mean s.d.i to the mean Ci (Ichise et al, 2002), resulting in 5.32%. Its value is similar to the mean s.d. across ROIs, 5.05%, estimated from the absolute deviation given by the residuals of the 2-CM fitting of TACs.

Statistics

Goodness of fit was evaluated using the Akaike Information Criterion (Akaike, 1974) and the Model Selection Criterion (MicroMath, 1995). Lower Akaike Information Criterion and higher Model Selection Criterion values were indicative of a better fit. Group data are expressed as mean±s.d.

Results

Safety Measures

On the basis of patient reports, electrocardiogram, blood pressure, and pulse, the injection of [18F]-FEPPA caused no adverse effects during the 3-hour scans. In addition, no significant effects were noted in any of the blood and urine tests acquired about 3±2 days after radioligand injection. The injected mass and activity of [18F]-FEPPA ranged from 0.12 to 3.22 μg and from 3.95 to 5.12 mCi, respectively (Supplementary Table 1).

Plasma Analysis

An average curve of unmetabolized [18F]-FEPPA in plasma (Figure 1A) reaches a maximum of 20.8 standard uptake value (SUV) at 23 seconds after injection and thereafter rapidly declines. Triexponential fitting describes very well (r2>0.99) the washout of radioactivity with half-lives of 0.06, 1.2, and 40 minutes. These half-lives are responsible for 8%, 23%, and 69% of the total area under the decay from the peak to infinity.

Figure 1.

Figure 1

Open in a new tab

(A) Average (n=12) time evolution of radioactivity in blood and unmetabolized [18F]-FEPPA in plasma. The inner plot shows details of the first 100 seconds postinjection. (B) Reverse-phase radiochromatogram of arterial plasma from one subject, 30 minutes postinjection of [18F]-FEPPA. Hilton method of analysis used with column switching at 3 minutes (Hilton et al, 2000). [18F]-FEPPA has a retention time of 9 minutes. (C) The average (n=12) percentage composition of plasma radioactivity over time is shown for [18F]-FEPPA (circles) and the radiometabolites (triangles, diamonds, and squares). SUV, standardized uptake value.

Reversed-phase high-performance liquid chromatography showed the presence of at least three radioactive metabolites of [18F]-FEPPA (Figure 1B), which appeared quickly in plasma and later became the predominant components (Figure 1C). The radiometabolites eluted earlier than did [18F]-FEPPA, indicating that they were less lipophilic than [18F]-FEPPA with a similar profile to that observed in rat (Wilson et al, 2008). [18F]-FEPPA was rapidly metabolized with ∼80% of the radioactivity in plasma attributable to polar metabolites after 30 minutes. The fraction of unmetabolized [18F]-FEPPA in plasma decreased more slowly thereafter, being 15%, 10%, 6.6%, 4.9%, and 3.6% at 45, 60, 90, 130, and 175 minutes postinjection, respectively (Figure 1C, circles). A Hill function (Gunn et al, 1998): 100% × (1−(atb/(cb+tb))) fitted the measure values very well (r2>0.997).

Distribution of Radioactivity in the Brain Regions

After [18F]-FEPPA injection, all subjects showed moderate levels of brain radioactivity that washed out gradually. The characteristics of the TACs are dependent on the ROI (Figure 2). The peak is very flat so it is difficult to identify with precision. It is reached within the first 20 minutes for all the gray matter TACs. Time–activity curves also present a sharp first peak immediately after injection from intravascular radioactivity. Thalamus shows the highest peak (1.45 SUV) followed by the cerebellum (1.4 SUV), temporal (1.35 SUV), occipital (1.33 SUV), frontal (1.3 SUV), putamen (1.29 SUV), caudate (1.28 SUV), insula (1.24 SUV), and cingulate (1.16 SUV). The washout is faster in the caudate, putamen, and cerebellum and slower in the thalamus. After 3 hours, the activity decreased to 52% of the peak value for the caudate, 53% for putamen, 57% for the cerebellum, between 63% and 64% for the cortical regions, 67% for thalamus. The white matter shows a lower peak (0.51 SUV) than all the other ROIs, reaching the peak later, at 54 minutes, and decaying slower than the thalamus to the 91% of the peak value at 3 hours. As expected from the known distribution of TSPOs in the human brain (Doble et al, 1987), the distribution of activity was widespread and fairly uniform within the gray matter of the cerebral cortices, cerebellum, basal ganglia, and thalamus (Supplementary Figure 1). As no brain region lacks TSPO expression, we could not apply a reference region method for the kinetic analysis.

Figure 2.

Figure 2

Open in a new tab

Average time–activity curve (TAC) (SUV, standardized uptake value; n=12) for thalamus, putamen, temporal cortex, and white matter.

Fluorine-labeled radioligands may be metabolized by defluorination, with subsequent uptake of 18F-fluoride ion into bone, including the skull. High levels of radioactivity in the skull could affect the quantification of, for example, cortical ROIs as consequence of spillover. However, radioactivity in bone was negligible (Supplementary Figure 1).

Kinetic Analysis

We compared 1-CM and unconstrained 2-CM by studying the results obtained from the full 3-hour session. For each ROI of each subject, the unconstrained 2-CM model provided a better fit than did the 1-CM (Figure 3), showing a lower Akaike Information Criterion and a higher Model Selection Criterion (Table 1).

Figure 3.

Figure 3

Open in a new tab

Time–activity data and curve fitting for occipital cortex for a typical subject. Two-tissue compartment model (2-CM) provided significantly better fitting than did 1-CM for all subjects. TAC, time–activity curve; SUV, standardized uptake value.

Table 1. AIC and MSC for 1-CM and 2-CM.

  1-CM
 2-CM
 P (paired t-test AIC) P (paired t-test MSC)
  AIC MSC AIC MSC    
Insula 106±9 0.6±0.3 33±29 2.2±0.5 2E–07 2E–07
Occipital cortex 106±6 0.6±0.3 1±19 2.9±0.4 3E–10 3E–10
Cerebellum 115±8 0.6±0.3 17±14 2.8±0.4 9E–11 8E–11
Temporal cortex 108±8 0.5±0.3 8±21 2.8±0.5 1E–09 1E–09
Frontal cortex 102±8 0.7±0.3 −9±22 3.2±0.5 1E–10 2E–10
Caudate 121±8 0.9±0.2 87±19 1.6±0.4 1E–05 8E–06
Putamen 117±11 0.8±0.3 64±25 2.0±0.5 2E–06 1E–06
Thalamus 120±8 0.3±0.2 51±25 1.8±0.6 8E–08 6E–08
Ant cingulate 118±14 0.4±0.3 63±35 1.6±0.6 9E–06 9E–06
White matter 117±15 0.1±0.4 65±32 1.2±0.6 5E–05 4E–05
Open in a new tab

AIC, Akaike Information Criterion; MSC, Model Selection Criterion; 1-CM, one-tissue compartment model; 2-CM, two-tissue compartment model.

Results obtained with the 2-CM are summarized in Table 2 (and Supplementary Figures 5–8). Excluding the white matter, the estimations of VT and VS with the unconstrained 2-CM and 3 hours of scanning present very good identifiability for all the ROIs (2% < %COV VT <7%, 2% < %COV VS <8%). The identifiability of BPND was excellent for large ROIs (e.g., frontal cortex 3.7%) and poorer for smaller ROIs (e.g., anterior cingulate 10% or caudate 12.5%). The rank order of VT is thalamus>temporal cortex>occipital cortex>insula>frontal>cerebellum>cingulated>putamen>caudate. The rank order is preserved for VS and BPND except for the anterior cingulate. While K1 (0.18±0.02) and K1/k2 (1.98±.21) are identifiable, k3 and k4 are not identifiable independently. The TAC of the white matter shows a completely different behavior to the other ROIs with a lower peak and a slower washout. Two-tissue compartment model describes the kinetic better than 1-CM. The identifiability of the parameters is lower than those of the gray matter: %COV=11.3%, 12.3%, and 21% for VT, VS, and BPND, respectively.

Table 2. Kinetic rate constants estimated with unconstrained 2-CM with 3 hours (above) and 2 hours (below) of scan data.

Region K1 COV K1/k2 COV k3 COV k4 COV VS COV VT COV k3/k4 COV
3 hours mL/cm3 per minute (%) mL/cm3 (%) 1/min (%) 1/min (%) mL/cm3 (%) mL/cm3 (%)   (%)
 
 Caudate 0.17±0.05 5.6±2.1 1.99±0.87 5.4±3.5 0.07±0.03 21±12 0.017±0.003 23±10 6.80±1.80 7.7±2.3 8.79±2.24 5.8±1.6 3.95±1.83 10±4
 Putamen 0.18±0.06 5.3±2.6 2.02±0.69 4.5±4.6 0.06±0.02 14±8 0.015±0.003 18±10 7.58±2.36 6.1±2.3 9.60±2.67 4.9±1.7 4.07±1.54 7.7±4.4
 Ant cingulate 0.15±0.04 5.9±4.1 1.74±0.56 6.3±5.0 0.07±0.02 12±6 0.012±0.004 17±5 9.61±3.24 7.5±2.8 11.35±3.59 6.5±2.4 5.67±1.67 9.9±4.1
 Cerebellum 0.20±0.06 2.8±1.7 2.06±0.69 2.9±2.5 0.06±0.02 8±3 0.013±0.001 10±2 9.32±2.74 3.6±0.7 11.38±2.98 3.0±0.5 4.92±1.81 5.0±2.1
 Frontal cx. 0.17±0.05 2.6±1.0 2.04±0.70 2.0±1.7 0.06±0.02 6±3 0.013±0.003 7±2 9.75±2.79 3.0±0.9 11.79±3.14 2.5±0.7 5.14±1.66 3.7±1.6
 Insula 0.16±0.04 3.2±1.1 2.00±0.63 4.1±3.2 0.06±0.02 10±3 0.012±0.003 13±3 9.91±2.92 5.1±1.4 11.91±3.26 4.3±1.1 5.18±1.69 7.0±2.5
 Occipital cx. 0.18±0.05 2.9±1.1 1.96±0.68 2.8±3.0 0.07±0.02 7±3 0.012±0.001 8±2 10.26±3.12 3.3±0.6 12.22±3.43 2.7±0.5 5.64±1.85 4.8±2.6
 Temporal cx. 0.19±0.05 2.6±1.0 2.03±0.71 2.2±1.5 0.07±0.03 8±3 0.012±0.002 10±3 10.38±3.09 3.6±0.8 12.41±3.40 3.0±0.7 5.65±2.47 4.4±1.1
 Thalamus 0.21±0.05 5.5±2.8 1.99±0.60 4.6±2.6 0.07±0.02 14±7 0.011±0.001 16±5 11.84±2.98 5.9±1.5 13.83±3.35 5.0±1.2 6.23±1.49 8.1±2.6
 White matter 0.07±0.03 7.2±3.4 0.51±0.13 15±12 0.10±0.03 22±19 0.007±0.004 25±13 7.56±3.02 12±5 8.07±3.09 11±4 152±5.8 21±12
2 hours
 Caudate 0.17±0.05 7.7±5.0 1.68±0.54 6.8±7.0 0.09±0.03 29±19 0.021±0.006 31±16 6.59±1.85 11±5 8.27±1.97 8.7±3.9 4.37±2.01 14±7
 Putamen 0.18±0.07 5.8±3.0 1.77±0.59 5.7±5.7 0.08±0.03 19±10 0.018±0.005 24±11 7.41±2.31 9.9±4.1 9.18±2.57 7.9±3.3 4.54±1.76 12±6
 Ant cingulate 0.16±0.04 5.6±3.0 1.57±0.56 5.3±4.9 0.08±0.03 18±13 0.014±0.004 27±12 9.08±2.84 13±5 10.65±3.12 11±4 6.14±2.22 14±6
 Cerebellum 0.21±0.06 2.7±1.3 1.79±0.53 3.3±2.3 0.08±0.03 8±3 0.015±0.002 12±3 8.77±2.56 5.5±1.7 10.55±2.66 4.5±1.4 5.36±2.44 6.9±2.4
 Frontal cx. 0.18±0.05 2.3±1.0 1.78±0.52 2.6±1.7 0.08±0.02 9±4 0.015±0.003 12±5 9.15±2.46 4.7±1.5 10.93±2.70 3.9±1.1 5.43±1.84 5.8±2.0
 Insula 0.17±0.04 4.1±1.5 1.82±0.43 4.3±2.8 0.07±0.01 13±9 0.014±0.003 21±11 9.47±2.93 9.1±3.1 11.29±3.05 7.5±2.5 5.39±1.98 10±4
 Occipital cx. 0.19±0.06 4.1±4.6 1.68±0.63 5.7±4.3 0.09±0.04 10±5 0.015±0.002 12±4 9.44±2.80 4.7±0.8 11.12±3.07 4.0±0.7 6.36±3.01 8.3±3.9
 Temporal cx. 0.19±0.05 3.2±1.2 1.76±0.60 2.9±2.6 0.09±0.04 9±5 0.014±0.002 13±5 9.69±2.65 5.7±1.5 11.46±2.93 4.8±1.3 6.17±2.98 6.8±2.1
 Thalamus 0.22±0.06 7.1±5.3 1.72±0.56 4.6±3.0 0.09±0.03 12±3 0.013±0.002 18±4 10.84±2.65 9.0±2.4 12.56±2.92 7.8±2.1 6.79±2.31 10±3
 White matter 0.07±0.03 8.4±4.4 0.53±0.14 19±16 0.10±0.05 39±54 0.009±0.005 50±43 6.70±2.68 23±12 7.23±2.76 21±11 12.9±5.3 30±15
Open in a new tab

COV, coefficient of variation; 2-CM, two-tissue compartment model.

The values are shown as mean±s.d. (n=12).

When the length of scanning is reduced from 3 to 2 hours, the rank order is mostly maintained (Table 2). Excluding the white matter, the %COV increases on average 2.5%, 3%, and 3.2% for VT, VS, and BPND, respectively. So while VT and VS continue to be reliably identifiable, only large ROIs give a BPND with good identifiability (e.g., frontal cortex %COV=5.8%). The correlation of values between 2 and 3 hours are excellent (r2=0.989 for VT, r2=0.994 for VS, and r2=0.971 for BPND) (Supplementary Figure 2). VS and VT, determined by 2 hours scanning, underestimates the values compared with 3 hours scanning by 6% and 7.5%, respectively. However, the spread of the measurements (s.d./means) are still practically the same (29% for VS and 27% for VT average across ROIs), suggesting no increased spread with the 2-hour scan. The measurements for white matter behave in the same way. Despite the identifiability being very low (%COV>21%), the correlation VT (and VS) between 2 and 3 hours is excellent (r2>0.98 in both cases).

Identifiability becomes poorer with a further reduction of length of scanning to 90 minutes. The average %COV excluding white matter are 14%, 17%, 19% for VT, VS, and BPND, respectively. The average of VT and VS does not present significant bias with respect to data from 2 hours scanning and the variability increases to 33% for VS and 31% for VT averaged across ROIs (Supplementary Table 2). The VT values at 90 minutes correlate with the values at 3 hours (r2=0.91, VT180 minutes=0.92, VT90 minutes+1.61).

Linear Graphical Analysis

VT values estimated with linear graphical approach using total least-squared method (VTTLS) (Varga and Szabo, 2002) correlate very well with those obtained using the 2-CM. For the 3-hour scans the relation is VTTLS=0.967 VT2-CM+0.0027, r2=0.970 and for the 2-hour scans, VTTLS=0.96 VT2-CM+0.08, r2=0.966 (Supplementary Figure 3). The mean underestimation of the linear approach is 3% for the 3-hour and 5% for the 2-hour scan.

Simulations

Change in k3 to simulate increase neuroinflammation

Table 3 (and Supplementary Figure 4) shows the result of the simulation using 2 hours scanning data. The results show that when k3 increases by a factor of six, the %COV for VT and VS increase from around 6% to around 10%, while the %COV BPND increases from around 10% to 19%, suggesting that VT and VS will be adequate binding parameters under these conditions.

Table 3. Simulation of increase in TSPO changing k3 and using a 2-CM with 2 hours of scan data.
Factor k3 Increase TSPO (%) VS %COV VS VT %COV VT k3/k4 %COV k3/k4
1 0 12.6±0.6 6.2±1.4 14.2±0.6 5.7±1.2 7.9±1.4 10.4±3.4
2 100 25.2±1.3 6.3±1.3 26.8±1.4 6.0±1.2 16.4±4.7 14.1±7.4
3 200 38.0±2.4 7.1±1.3 39.7±2.4 6.8±1.3 23.2±7.7 14.3±8.2
4 300 50.5±0.6 8.0±1.4 52.2±0.6 7.8±1.2 29.9±1.4 14.7±3.4
5 400 63.4±5.1 9.2±1.7 65.1±5.0 8.9±1.7 37.4±7.0 17.7±11.8
6 500 76.3±6.8 10.3±2.0 78.1±6.7 10.1±1.9 43.3±13.9 19.1±14.1
Open in a new tab

COV, coefficient of variation; TSPO, translocator protein 18 kDa; VS, specific distribution volume; VT, total distribution volume; 2-CM, two-tissue compartment model.

Discussion

This work describes the first human quantification of [18F]-FEPPA for the in vivo estimation of TSPO density in human brain. The present results indicate that the 2-CM always describes the kinetics of [18F]-FEPPA better than 1-CM and that VT, VS, and BPND can be estimated with a reasonably good identifiability. The rank order of [18F]-FEPPA VT, VS, and BPND is thalamus>temporal cortex>occipital cortex>insula>frontal>cerebellum>cingulated>putamen>caudate.

While scanning for 3 hours looks optimum from the point of view of identifiability, it has some practical limitations in human PET studies since some subjects do not tolerate scanning beyond 2 hours. Shortening the length of the scan is possible but with somewhat poorer identifiably for BPND, although identifiability of VT and VS are preserved (averaging across ROIs: %COV increases from 5% to 8%, and from 4% to 7% for VS and VT, respectively). Additionally, a bias is identified in the magnitudes. In the average across the ROIs, VS values decrease 6% and VT values decrease 7.5%. The bias depends on the ROIs, with the putamen presenting the smallest bias and the thalamus the biggest bias. Nevertheless, the bias is well below 10%, making it quite acceptable for clinical studies.

We found a large variability of VT, VS, and BPND across the 12 subjects. Depending on the ROI for 2 hours of scan data, the variability of VT ranges between 23% and 29%, of VS between 24% and 31%, and of BPND between 34% and 48%. This variability is, however, similar to the one observed with [18F]-PBR06 (VT 26% to 34%) (Fujimura et al, 2009). The variability in our study was not a consequence of gender or age. While there was no correlation with either gender or age, there are not enough data to reach a conclusion at this point. In vitro studies (Owen et al, 2010, 2011) have shown two TSPO binding sites for [11C]-PBR28, which have lead to the classification of the population in three groups: low affinity binders, mixed affinity binder, and high affinity binders. In the same study, in vitro results are not incompatible with the existence of two binding sites for other TSPO radioligands (e.g., [18F]-PBR06, [11C]-DAA1106, [11C]-DPA713, and [11C]-PBR111). However, in vivo low affinity binders (∼15% of the population) was only reported with [11C]-PBR28 (Fujita et al, 2008), and it has not been possible to determine with certainty the presence of mixed affinity binder and/or high affinity binders in a sample of 37 subjects (Owen et al, 2011). We have not found so far nonbinders (i.e., low affinity binders) with [18F]-FEPPA. On the other hand with only 12 subjects, it is not possible to classify our sample into groups (i.e., low affinity binders, mixed affinity binder, and high affinity binders). The possibility that the variability in the results could be due to the existence of different groups characterized by the binding sites cannot be ruled out.

Reducing the length of scanning does not change this variability and, importantly, our simulations predict that with 2 hours of scanning, we will still be able to use [18F]-FEPPA to measure increases up to 500% in TSPO densities. It should be noted that k3 would only change linearly with TSPO densities (Bmax) if both the free fraction of ligand in the nondisplaceable tissue compartment (fND) and the association rate constant (kon) does not change (k3=fND kon Bmax). Eventually, under pathological conditions characterized by a higher Bmax those variables and K1, k2, and k4 could change independently. Simulation showed that changes in blood flow would not affect the quantification of VT and VS (simulations described in Supplementary Information).

The Logan graphical method using total least-squared approach (Varga and Szabo, 2002) correlates well with the 2-CM so it would be a good option for parametric maps of VT. While VT is definitively a good option for ROI analysis, VS may be better as it does not include free and nonspecific contributions and its identifiability is very robust. However, it should be noted that optimization algorithms are prone to finding solutions (local minimum of the cost function) that in spite of presenting good identifiability could provide estimations of K1/k2 that may not represent the physiological volume of distribution of the free and nonspecific compartment (VND). In the absence of some comparisons of VT between conditions of blocked and unblocked specific binding, it is not possible to know the true decomposition of VT into VND+VS. In this work, we could easily distinguish between compartments, and regional estimates of K1/k2, which were consistently between 1.5 and 2 in all brain regions, but this is still not direct evidence that these values represent the physiological VND.

Conclusion

Binding of [18F]-FEPPA in the healthy human brain was well identified with an unconstrained 2-CM using 120 minutes of scan data. VT and VS are adequate binding parameters to quantify [18F]-FEPPA in living humans with increases of TSPO until up 500%. In summary, [18F]-FEPPA is a promising PET tool to measure neuroinflammation in human brain, demonstrating an excellent and feasible method for obtaining VT, a useful index of TSPO binding in living humans.

Acknowledgments

The authors thank the staff of the PET Centre for the acquisition of data, Thiviya Selvanathan and Sharon Hung for helping to reproduce the results while learning PET analysis, Dr Isabelle Boileau for her critical revision of the manuscript, and the scientists that came to the poster during the Neuroreceptor Mapping conference in Glasgow for their insightful comments on the work. The authors acknowledge the anonymous reviewers for their thoughtful suggestions and comments. This work has been partially supported by the Scottish Rite Charitable Foundation of Canada, Canada Foundation for Innovation (CFI) and the Ontario Ministry of Research and Innovation.

The authors declare no conflict of interest.

Footnotes

Supplementary Information accompanies the paper on the Journal of Cerebral Blood Flow & Metabolism website (http://www.nature.com/jcbfm)

Supplementary Material

Supplementary Information
Click here for additional data file. (791.5KB, doc)

References

  1. Akaike H. A new look at the statistical model identification. IEEE Trans Automat Contr. 1974;19:716–723. [Google Scholar]
  2. Bencherif B, Stumpf MJ, Links JM, Frost JJ. Application of MRI-based partial-volume correction to the analysis of PET images of mu-opioid receptors using statistical parametric mapping. J Nucl Med. 2004;45:402–408. [PubMed] [Google Scholar]
  3. Bennacef I, Salinas C, Horvath G, Gunn R, Bonasera T, Wilson A, Gee A, Laruelle M. Comparison of [11C]PBR28 and [18F]FEPPA as CNS peripheral benzodiazepine receptor PET ligands in the pig. J Nucl Med Meeting Abstracts. 2008;49:81P–81b. [Google Scholar]
  4. Bernassau JM, Reversat JL, Ferrara P, Caput D, Lefur G.1993A 3D model of the peripheral benzodiazepine receptor and its implication in intra mitochondrial cholesterol transport J Mol Graph 11236–244.235 [DOI] [PubMed] [Google Scholar]
  5. Boutin H, Chauveau F, Thominiaux C, Gregoire MC, James ML, Trebossen R, Hantraye P, Dolle F, Tavitian B, Kassiou M. [11C]-DPA-713: a novel peripheral benzodiazepine receptor PET ligand for in vivo imaging of neuroinflammation. J Nucl Med. 2007;48:573–581. doi: 10.2967/jnumed.106.036764. [DOI] [PubMed] [Google Scholar]
  6. Braestrup C, Albrechtsen R, Squires RF. High densities of benzodiazepine receptors in human cortical areas. Nature. 1977;269:702–704. doi: 10.1038/269702a0. [DOI] [PubMed] [Google Scholar]
  7. Burger C, Buck A. Requirements and implementation of a flexible kinetic modeling tool. J Nucl Med. 1997;38:1818–1823. [PubMed] [Google Scholar]
  8. Cagnin A, Myers R, Gunn RN, Lawrence AD, Stevens T, Kreutzberg GW, Jones T, Banati RB. In vivo visualization of activated glia by [11C] (R)-PK11195-PET following herpes encephalitis reveals projected neuronal damage beyond the primary focal lesion. Brain. 2001;124:2014–2027. doi: 10.1093/brain/124.10.2014. [DOI] [PubMed] [Google Scholar]
  9. Camsonne R, Crouzel C, Comar D, Mazière M, Prenant C, Sastre J, Moulin M, Syrota A. Synthesis of N-(11C)methyl, N-(methyl-1-propyl), (chloro-2-phenyl)-1-isoquinoleine carboxamide-3 (PK-11195) : a new ligand for peripheral benzodiazepine receptors. J Labelled Comp Radiopharm. 1984;21:985–991. [Google Scholar]
  10. Charbonneau P, Syrota A, Crouzel C, Valois JM, Prenant C, Crouzel M. Peripheral-type benzodiazepine receptors in the living heart characterized by positron emission tomography. Circulation. 1986;73:476–483. doi: 10.1161/01.cir.73.3.476. [DOI] [PubMed] [Google Scholar]
  11. Chen MK, Guilarte TR. Translocator protein 18 kDa (TSPO): molecular sensor of brain injury and repair. Pharmacol Ther. 2008;118:1–17. doi: 10.1016/j.pharmthera.2007.12.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Culty M, Li H, Boujrad N, Amri H, Vidic B, Bernassau JM, Reversat JL, Papadopoulos V. In vitro studies on the role of the peripheral-type benzodiazepine receptor in steroidogenesis. J Steroid Biochem Mol Biol. 1999;69:123–130. doi: 10.1016/s0960-0760(99)00056-4. [DOI] [PubMed] [Google Scholar]
  13. Debruyne JC, Versijpt J, Van Laere KJ, De Vos F, Keppens J, Strijckmans K, Achten E, Slegers G, Dierckx RA, Korf J, De Reuck JL. PET visualization of microglia in multiple sclerosis patients using [11C]-PK11195. Eur J Neurol. 2003;10:257–264. doi: 10.1046/j.1468-1331.2003.00571.x. [DOI] [PubMed] [Google Scholar]
  14. Doble A, Malgouris C, Daniel M, Daniel N, Imbault F, Basbaum A, Uzan A, Gueremy C, Le Fur G. Labelling of peripheral-type benzodiazepine binding sites in human brain with [3H]PK 11195: anatomical and subcellular distribution. Brain Res Bull. 1987;18:49–61. doi: 10.1016/0361-9230(87)90033-5. [DOI] [PubMed] [Google Scholar]
  15. Fujimura Y, Ikoma Y, Yasuno F, Suhara T, Ota M, Matsumoto R, Nozaki S, Takano A, Kosaka J, Zhang MR, Nakao R, Suzuki K, Kato N, Ito H. Quantitative analyses of 18F-FEDAA1106 binding to peripheral benzodiazepine receptors in living human brain. J Nucl Med. 2006;47:43–50. [PubMed] [Google Scholar]
  16. Fujimura Y, Zoghbi SS, Simeon FG, Taku A, Pike VW, Innis RB, Fujita M. Quantification of translocator protein (18 kDa) in the human brain with PET and a novel radioligand, 18F-PBR06. J Nucl Med. 2009;50:1047–1053. doi: 10.2967/jnumed.108.060186. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Fujita M, Imaizumi M, Zoghbi SS, Fujimura Y, Farris AG, Suhara T, Hong J, Pike VW, Innis RB. Kinetic analysis in healthy humans of a novel positron emission tomography radioligand to image the peripheral benzodiazepine receptor, a potential biomarker for inflammation. Neuroimage. 2008;40:43–52. doi: 10.1016/j.neuroimage.2007.11.011. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Gehrmann J, Matsumoto Y, Kreutzberg GW. Microglia: intrinsic immuneffector cell of the brain. Brain Res Rev. 1995;20:269–287. doi: 10.1016/0165-0173(94)00015-h. [DOI] [PubMed] [Google Scholar]
  19. Gerhard A, Banati RB, Goerres GB, Cagnin A, Myers R, Gunn RN, Turkheimer F, Good CD, Mathias CJ, Quinn N, Schwarz J, Brooks DJ. [11C](R)-PK11195 PET imaging of microglial activation in multiple system atrophy. Neurology. 2003;61:686–689. doi: 10.1212/01.wnl.0000078192.95645.e6. [DOI] [PubMed] [Google Scholar]
  20. Gerhard A, Neumaier B, Elitok E, Glatting G, Ries V, Tomczak R, Ludolph AC, Reske SN. In vivo imaging of activated microglia using [11C]PK11195 and positron emission tomography in patients after ischemic stroke. Neuroreport. 2000;11:2957–2960. doi: 10.1097/00001756-200009110-00025. [DOI] [PubMed] [Google Scholar]
  21. Gunn RN, Sargent PA, Bench CJ, Rabiner EA, Osman S, Pike VW, Hume SP, Grasby PM, Lammertsma AA. Tracer kinetic modeling of the 5-HT1A receptor ligand [carbonyl-11C]WAY-100635 for PET. Neuroimage. 1998;8:426–440. doi: 10.1006/nimg.1998.0379. [DOI] [PubMed] [Google Scholar]
  22. Hatano K, Yamada T, Toyama H, Kudo G, Nomura M, Suzuki H, Ichise M, Wilson AA, Sawada M, Kato T, Ito K. Correlation between FEPPA uptake and microglia activation in 6-OHDA injured rat brain. NeuroImage. 2010;52:S138–S13S. [Google Scholar]
  23. Hilton J, Yokoi F, Dannals RF, Ravert HT, Szabo Z, Wong DF. Column-switching HPLC for the analysis of plasma in PET imaging studies. Nucl Med Biol. 2000;27:627–630. doi: 10.1016/s0969-8051(00)00125-6. [DOI] [PubMed] [Google Scholar]
  24. Ichise M, Toyama H, Innis RB, Carson RE. Strategies to improve neuroreceptor parameter estimation by linear regression analysis. J Cereb Blood Flow Metab. 2002;22:1271–1281. doi: 10.1097/01.WCB.0000038000.34930.4E. [DOI] [PubMed] [Google Scholar]
  25. Iida H, Kanno I, Miura S, Murakami M, Takahashi K, Uemura K. Error analysis of a quantitative cerebral blood flow measurement using H2(15)O autoradiography and positron emission tomography, with respect to the dispersion of the input function. J Cereb Blood Flow Metab. 1986;6:536–545. doi: 10.1038/jcbfm.1986.99. [DOI] [PubMed] [Google Scholar]
  26. Innis RB, Carson R. Consensus nomenclature: its time has come. Eur J Nucl Med Mol Imaging. 2007;34:1239. doi: 10.1007/s00259-007-0481-7. [DOI] [PubMed] [Google Scholar]
  27. James ML, Selleri S, Kassiou M. Development of ligands for the peripheral benzodiazepine receptor. Curr Med Chem. 2006;13:1991–2001. doi: 10.2174/092986706777584979. [DOI] [PubMed] [Google Scholar]
  28. Kreutzberg GW. Microglia: a sensor for pathological events in the CNS. Trends Neurosci. 1996;19:312–318. doi: 10.1016/0166-2236(96)10049-7. [DOI] [PubMed] [Google Scholar]
  29. Leenders KL, Perani D, Lammertsma AA, Heather JD, Buckingham P, Healy MJ, Gibbs JM, Wise RJ, Hatazawa J, Herold S, Beaney RP, Brooks DJ, Spinks T, C R, Frackowiak R, Jones T. Cerebral blood flow, blood volume and oxygen utilization. Normal values and effect of age. Brain. 1990;113 (Part 1:27–47. doi: 10.1093/brain/113.1.27. [DOI] [PubMed] [Google Scholar]
  30. Logan J, Fowler JS, Volkow ND, Ding YS, Wang GJ, Alexoff DL. A strategy for removing the bias in the graphical analysis method. J Cereb Blood Flow Metab. 2001;21:307–320. doi: 10.1097/00004647-200103000-00014. [DOI] [PubMed] [Google Scholar]
  31. McEnery MW, Snowman AM, Trifiletti RR, Snyder SH. Isolation of the mitochondrial benzodiazepine receptor: association with the voltage-dependent anion channel and the adenine nucleotide carrier. Proc Natl Acad Sci USA. 1992;89:3170–3174. doi: 10.1073/pnas.89.8.3170. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Meyer E. Simultaneous correction for tracer arrival delay and dispersion in CBF measurements by the H215O autoradiographic method and dynamic PET. J Nucl Med. 1989;30:1069–1078. [PubMed] [Google Scholar]
  33. MicroMath . MicroMath Scientist Handbook Rev 7EEF. MicroMath: Salt Lake City; 1995. p. pp 467. [Google Scholar]
  34. Mintun MA, Raichle ME, Kilbourn MR, Wooten GF, Welch MJ. A quantitative model for the in vivo assessment of drug binding sites with positron emission tomography. Ann Neurol. 1984;15:217–227. doi: 10.1002/ana.410150302. [DOI] [PubMed] [Google Scholar]
  35. Okubo T, Yoshikawa R, Chaki S, Okuyama S, Nakazato A. Design, synthesis, and structure-activity relationships of novel tetracyclic compounds as peripheral benzodiazepine receptor ligands. Bioorg Med Chem. 2004;12:3569–3580. doi: 10.1016/j.bmc.2004.04.025. [DOI] [PubMed] [Google Scholar]
  36. Owen DR, Gunn RN, Rabiner EA, Bennacef I, Fujita M, Kreisl WC, Innis RB, Pike VW, Reynolds R, Matthews PM, Parker CA. Mixed-affinity binding in humans with 18-kDa translocator protein ligands. J Nucl Med. 2011;52:24–32. doi: 10.2967/jnumed.110.079459. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Owen DR, Howell OW, Tang SP, Wells LA, Bennacef I, Bergstrom M, Gunn RN, Rabiner EA, Wilkins MR, Reynolds R, Matthews PM, Parker CA. Two binding sites for [3H]PBR28 in human brain: implications for TSPO PET imaging of neuroinflammation. J Cereb Blood Flow Metab. 2010;30:1608–1618. doi: 10.1038/jcbfm.2010.63. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Papadopoulos V, Amri H, Boujrad N, Cascio C, Culty M, Garnier M, Hardwick M, Li H, Vidic B, Brown AS, Reversa JL, Bernassau JM, Drieu K. Peripheral benzodiazepine receptor in cholesterol transport and steroidogenesis. Steroids. 1997;62:21–28. doi: 10.1016/s0039-128x(96)00154-7. [DOI] [PubMed] [Google Scholar]
  39. Pappata S, Cornu P, Samson Y, Prenant C, Benavides J, Scatton B, Crouzel C, Hauw JJ, Syrota A. PET study of carbon-11-PK 11195 binding to peripheral type benzodiazepine sites in glioblastoma: a case report. J Nucl Med. 1991;32:1608–1610. [PubMed] [Google Scholar]
  40. Phelps ME, Huang SC, Hoffman EJ, Kuhl DE. Validation of tomographic measurement of cerebral blood volume with C-11-labeled carboxyhemoglobin. J Nucl Med. 1979;20:328–334. [PubMed] [Google Scholar]
  41. Rusjan P, Mamo D, Ginovart N, Hussey D, Vitcu I, Yasuno F, Tetsuya S, Houle S, Kapur S. An automated method for the extraction of regional data from PET images. Psychiatry Res. 2006;147:79–89. doi: 10.1016/j.pscychresns.2006.01.011. [DOI] [PubMed] [Google Scholar]
  42. Studholme C, Hill DLG, Hawkes DJ. An overlap invariant entropy measure of 3D medical image alignment. Pattern Recognition. 1999;32:71–86. [Google Scholar]
  43. Varga J, Szabo Z. Modified regression model for the Logan plot. J Cereb Blood Flow Metab. 2002;22:240–244. doi: 10.1097/00004647-200202000-00012. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Wilson AA, Garcia A, Parkes J, McCormick P, Stephenson KA, Houle S, Vasdev N. Radiosynthesis and initial evaluation of [18F]-FEPPA for PET imaging of peripheral benzodiazepine receptors. Nucl Med Biol. 2008;35:305–314. doi: 10.1016/j.nucmedbio.2007.12.009. [DOI] [PubMed] [Google Scholar]
  45. Yaqub M, Boellaard R, Kropholler MA, Lammertsma AA. Optimization algorithms and weighting factors for analysis of dynamic PET studies. Phys Med Biol. 2006;51:4217–4232. doi: 10.1088/0031-9155/51/17/007. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Information
Click here for additional data file. (791.5KB, doc)

Articles from Journal of Cerebral Blood Flow & Metabolism are provided here courtesy of SAGE Publications

RESOURCES