Download Manual Do Receptor Century Br 2014

Posted by admin
Published online 2018 Aug 2. doi: 10.1186/s13550-018-0416-2
PMID: 30069645

Associated Data

Supplementary Materials
Additional file 1: Table S1. Mean BPND values of [11C]SCH23390 in the morning and afternoon, the absolute variability, and ICC in 15 healthy men in four brain regions. There are eight BPND values for each region derived from the different combinations of methods in the analysis process. (DOCX 18 kb)
GUID: 9EA22F74-24EA-4F5E-A04E-26EDC9D60923
Data Availability Statement

Please contact the author for data requests.

Abstract

Background

The D1-dopamine receptor radioligand [11C]SCH23390 has been frequently used in PET studies. In drug-naïve patients with schizophrenia, the findings have been inconsistent, with decreases, increases, and no change in the frontal cortex D1-dopamine receptors. While these discrepancies are likely primarily due to a lack of statistical power in these studies, we speculated that an additional explanation may be the differences due to methods of image analysis between studies, affecting reliability as well as bias between groups.

Methods

Fifteen healthy subjects underwent two PET measurements with [11C]SCH23390 on the same day. The binding potential (BPND) was compared using a 95% confidence interval following manual and automated delineation of a region of interest (ROI) as well as with and without frame-by-frame realignment.

Results

Automated target region delineation produced lower BPND values, while automated delineation of the reference region yielded higher BPND values. However, no significant differences were observed for repeatability using automated and manual delineation methods. Frame-by-frame realignment generated higher BPND values and improved repeatability.

Conclusions

The results suggest that the choice of ROI delineation method is not an important factor for reliability, whereas the improved results following movement correction confirm its importance in PET image analysis. Realignment is therefore especially important for measurements in patient populations such as schizophrenia or Parkinson’s disease, where motion artifacts may be more prevalent.

Electronic supplementary material

The online version of this article (10.1186/s13550-018-0416-2) contains supplementary material, which is available to authorized users.

Keywords: Positron emission tomography, ROI, Realignment, FreeSurfer, D1-dopamine receptor

Background

The development of molecular imaging in the 1980s was largely driven by the need to examine the dopamine system in CNS disorders such as schizophrenia and Parkinson’s disease [, ]. For this reason, radioligands were developed for the D1-dopamine receptor (D1-DR), the D2-dopamine receptor (D2-DR), and for the presynaptic synthesis of dopamine. In the research on the pathophysiology of schizophrenia, a large number of PET studies on the dopamine system have since then been reported and covered in reviews and meta-analyses [, ]. However, for the dopamine receptors and transporters, the results have been inconsistent in several respects. One example is D1-DR binding, which has been reported as lower, unchanged, and higher in the frontal cortex [–]. Medication effects may explain some of the differences between the results since experimental studies have shown that D1-R expression is affected by antipsychotic medication [–]. However, studies in drug-naïve patients have also been inconclusive [–]. For these analyses, sample sizes have typically been small, leading to an increased risk for both type I and type II errors [, , ]. A wider consideration is that the series of D1-DR studies in schizophrenia has been reported over many years during which time the methods for image analysis have been improved, including the development of new software and strategies for the definition of regions of interest (ROI) and methods for motion correction. Hence, it cannot be excluded that differences in methodological reliability may also contribute to the discrepant findings to some degree.

In the early years of PET imaging, the delineation of ROIs was performed manually directly on the PET images and guided only by the regional distribution of radioactivity. In the late 1980s, the advent of magnetic resonance imaging (MRI) allowed for manual ROI definition on individual MR images and co-registration to the PET images. Subsequently, structural brain atlases and software tools were developed for automated definition of ROIs, offering advantages in terms of reduced investigator bias and workload [].

The agreement between manual and automated ROI delineations on PET-measured receptor-binding measurements using PET has been addressed in several studies. Good agreement has been demonstrated for tracers with widespread homogenous cortical binding [–] as well as more restricted subcortical binding [, ]. For [11C] raclopride binding to the D2-DR, which is restricted to the striatum, there was also a good agreement in repeatability as assessed using a test-retest design []. In contrast to the D2-DR, the D1-DR is widely expressed in the cortical regions. Importantly, the proximity of the cortical regions to the subarachnoidal space provides specific challenges for ROI definition and may also render measurements more susceptible to motion artifacts.

The primary aim of the present PET study with [11C]SCH23390 in healthy subjects was to compare manual and automated methods for ROI definition of both subcortical and cortical regions in a test-retest design. The secondary aim was to evaluate the effect of motion correction, and the third was to compare the use of manual and automated ROIs for the reference region, cerebellum. The methods were compared with regard to the size and spatial agreement of the ROIs, the binding potential (BPND) values, and repeatability of the BPND values.

Methods

Sixteen healthy subjects, age 22–35 years, were enrolled at the Center for Psychiatry Research, Department of Clinical Neuroscience, Karolinska Institutet, and Stockholm County Council, Stockholm, Sweden. One subject was excluded from the analysis due to incorrect head positioning in the PET system rendering a small portion of the cerebellum visible in only one image section. All subjects were male since this study was made in preparation for a planned phase 1 trial on drug-induced D1-DR occupancy.

The subjects were healthy according to medical history, physical examination, MRI, blood and urine chemistry, and psychiatric screening based on M.I.N.I. interview, Becks Anxiety Inventory, the Montgomery-Åsberg Depression Rating Scale, and the Alcohol Use Disorders Identification Test (AUDIT). They had no history of alcohol or drug addiction or abuse [], frequent nicotine use, or history or presence of epilepsy and brain injury. A urine drug screening was performed at screening and before each PET measurement.

All subjects underwent MRI, performed on a 1.5-T Siemens Avanto imaging system (Siemens AG, Muenchen, Germany) at Praktikertjänst Röntgen, Odenplan, Stockholm, Sweden. Each individual underwent one MR examination. A T2-weighted measurement was performed to rule out any brain abnormality, and a T1-weighted measurement was performed with isometric 1-mm voxels used for gray and white matter segmentation and delineation of ROIs. The T2 protocol used the following sequence: repetition time/echo time = 4990/100 ms, field of view 230 mm, image matrix 18 blades, flip angle 150°, and slice thickness = 5 mm. The T1 protocol used a 3D sagittal magnetization-prepared rapid gradient-echo (3D MP-RAGE) with the following sequence: repetition time/echo time/inversion time = 1790/3.53/1100 ms, field of view 260 mm, image matrix 256 mm × 208 mm, flip angle 15°, and slice thickness = 1 mm.

For each subject, examinations were performed twice on the ECAT EXACT HR PET system, in the morning and afternoon, respectively. Radioactivity in the brain was measured with 3D data acquisition. The spatial resolution in the reconstructed sections is 3.6 mm at the center of the field of view []. A transmission scan was performed using three rotating 68Ge rod sources for 5 min.

To minimize head movement during the PET measurement, a plaster helmet was made for each subject individually, for use during the PET measurement []. At the start of the PET measurement, a sterile phosphate buffer (pH = 7.4) containing [11C]SCH23390 was injected as a bolus during several seconds into the cubital vein. The venous catheter was then immediately flushed with up to 10 mL saline solution.

The injected radioactivity was 333 ± 44 and 332 ± 49 MBq (mean, SD), in the morning and the afternoon, respectively, with a specific radioactivity of 507 ± 244 and 574 ± 238 MBq/nmol, which correspond to a mass of 0.23 ± 0.13 μg and 0.20 ± 0.10 μg (mean ± SD) [].

Following injection, emission data were collected for 51 min in a sequence of 13 time frames. The time frames of acquisition data were reconstructed into a series of 3D PET images of radioactivity concentration. Images were reconstructed to correct for attenuation and scatter using 2D filtered back projection, with a Hanning filter (2.0 mm) on a 128 × 128 matrix and azoom = 2.17. The voxel size was 2.030 × 2.030 × 3.125 mm.

MR and PET image processing

Processing of MRI and PET data

The PET images were corrected for head movement using a frame-by-frame realignment algorithm, in which all frames were individually realigned to the first minute of acquisition [] using SPM5 (Wellcome Department of Cognitive Neurology, University College London) [27]. For comparison, a parallel data set was created for which no realignment was performed.

The T1-weighted MR images were reoriented according to the line defined by the anterior and posterior commissures (AC-PC line) being parallel to the horizontal plane and the inter-hemispheric plane parallel to the sagittal plane. The MR images were then co-registered to the summation PET image (9–51 min) using SPM5 (Wellcome Department of Imaging Neuroscience, London, UK) using the normalized mutual information algorithm [28] and the default 7 × 7 FWHM smoothing of the 256 × 256 joint histogram.

The MR images were used to delineate anatomical regions of interest (ROI) for the caudate nucleus (CAU), the putamen (PUT), the dorsolateral prefrontal cortex (DLPFC), and the insular cortex (INS). Each region was defined by a manual and an automated method, respectively. The regions were chosen to represent the regions of central interest in schizophrenia research and include high- and low D1-DR density regions with different degrees of proximity to CSF. Regions with more CSF borders may be particularly vulnerable to errors in ROI definition or motion artifacts. The caudate and putamen are high-D1-DR density regions bordering and not bordering to CSF, respectively. The DLPFC and INS are low-density D1-DR regions bordering more and less to CSF, respectively.

Manual ROIs

For the manual method, an in-house software, HBA [], was used where the reoriented MP-RAGE volume was loaded for manual delineation of the ROIs on any of the three orthogonal projections. The manual segmentation was performed by one investigator (PS) who has more than 20 years of experience in manual ROI delineation. The CAU and PUT were delineated as described by Mawlawi et al. [], with the modification that the sagittal planes were used instead of the coronal. The DLPFC was traced on all the coronal planes anterior to the genu of the corpus callosum. The INS was delineated on all of its transaxial planes. The cerebellum was drawn on the central six transaxial images of the cerebellum separately on each hemisphere and about 1 cm distant from the subarachnoidal space. The manual ROIs were not masked by the GM map in order to have a complete manual method of gray-white matter segmentation for comparison with the automated ROIs that were masked by the GM map. The ROIs were translated into each PET study space using the inter-modality co-registration matrices.

Automated ROIs

The automated definition of target ROIs was performed using FreeSurfer (FS, version 5.0.0, http://surfer.nmr.mgh.harvard.edu/) [] to obtain subject-specific anatomical delineation by reconstruction of the cortex and segmentation of subcortical structures as described elsewhere [, ]. The FreeSurfer morphometric procedures have been shown to exhibit good reproducibility across scanner manufacturers and across different field strengths [, ] and have been validated against histological [] and manual measurements []. In addition, the cortical structures are divided based on individual cortical folding patterns to match the cortical geometry across subjects [].

The ROI for the dorsolateral prefrontal cortex used here comprises the ctx-rostralmiddlefrontal and ctx-parstriangularis regions of the Desikan-Killiany Atlas in FreeSurfer [].

Finally, an automated ROI for the cerebellum was defined using FSL (the FMRIB Software Library), a library of analysis tools for brain imaging data []. This region was defined using the maximum probability FSL MNIfnirt atlas segmentation with 25% probability threshold [], from which a mask was defined containing cerebellar regions VI, crus I, and crus II (indices [, , , , , ]). This mask was then registered to the space of each individual’s MR using the inverse FNIRT warp parameters. This mask was subsequently then trimmed as follows: 8 mm from the cortex, 8 mm from the vermis (defined using the same atlas), 4 mm from the edge of the brain mask (FSL brain mask). In addition, voxels belonging to the two most inferior planes of the PET image were excluded from the ROI. The resulting ROI was multiplied by the FreeSurfer gray matter segmentation mask to obtain a ROI consisting of only those voxels identified as belonging to gray matter. For more information, see Matheson et al. []. The purpose was to compare the different impact on the BPND value using a standardized automatically delineated cerebellum compared to using the manual cerebellum as reference region.

Time-activity curves

The two ROI data sets (manual and automated) were applied to extract time-activity curves (TACs) from the four dynamic PET images (test and retest each processed with or without realignment) amounting to eight sets of TACs.

Calculation of BPND values

The regional BPND values for [11C]SCH23390 binding to D1-DR were calculated with the simplified reference tissue model using the cerebellum curve as an estimate for non-specifically bound [11C]SCH23390 []. The eight sets of TACs were analyzed twice using the manual and modified FSL cerebellum ROI-derived TAC, respectively, amounting to 16 sets of BPND values for each measurement.

Comparison of ROI volumes

To compare manually and automatically generated ROIs (FreeSurfer and FSL), the number of voxels for each paired ROI (both hemispheres) was extracted from the ROI files and converted to cubic millimeter. The spatial agreement between the methods was expressed by the Jaccard coefficient which is the ratio between the size of the intersection and the size of the union of the voxel sets:

1

The coefficient can vary between 0 (no agreement) and 1 (perfect match).

In addition, the spatial agreement was estimated for the relatively smaller manual ROI by the percentage of the manual ROI covered by the FreeSurfer or FSL ROI:

2

The values increase with the overlap of the manual and automated ROIs so that 100% indicates that the manual ROI is completely encompassed by the FreeSurfer or FSL ROI.

Statistics

Statistical analysis—general

In the following, the terminology related to the statistical analysis follows the recommendations of the Terminology Working Group of the Quantitative Imaging Biomarker Alliance []. The inclusion of a test-retest analysis serves two purposes. First, it allows for the examination of the repeatability of PET measurements with [11C]SCH23390. Second, it can be viewed as a duplicate measure of the same variable, i.e., in the same way as the mean of triplicate measurements commonly are used in biochemistry and bioanalyses.

Test-retest analysis

Repeatability was measured by calculating the absolute variability (AV) expressed as the difference in the BPND values between the first and second PET measurements relative to the mean of the two values according to the following equation:

3

The intraclass correlation coefficient (ICC) was used as a relative measure of reliability in BPND, where the variability between the test and retest BPND is related to the variability of BPND among the subjects according to the following equation:

4

where MSB is the between-subject mean sum of squares, MSW is the within-subject mean sum of squares, and k represents the number of observations (in this case 2). The ICC is an estimation of reliability, which is the “true” variance/total variance. Thus, a value of 1 means that there is no measurement error at all, while a value of 0 means that the obtained measurements are entirely comprised of measurement error. Zero is therefore technically the lowest value for the ICC, and a negative value can be approximated by 0.

Comparisons of ROI volumes and BPND values between the methods for image analyses

The advantages of interval estimate compared to point estimate as the output of statistical analysis have been demonstrated for medical research in general [, ] and specifically for the comparison of quantitative biomarkers []. Hypothesis testing for a point estimate using a p value is a binary (yes or no) decision even for minor differences that may not be of practical importance []. It may thus be more useful to evaluate the data with an interval estimate that in addition to a p value provides a plausible range for the true value, which can be judged subjectively on an agreement between and precision of the methods [].

Computation of median, interquartile range (25th to 75th), and coefficient of variation (CoV; standard deviationmean) was applied to evaluate the distribution and variability of the BPND values. Computation of a 95% confidence interval was applied for the evaluation of the agreement and precision of the two ROI volumes (manual and automated), and the two BPND values from the three comparisons in the image analysis (ROI definition, realignment and definition of reference region).

The two PET measurements and the two methods in each of the three comparisons of the analysis amounted to 2 × 2 × 2 × 2 equals 16 sets or 8 pairs of BPND values for each ROI as illustrated in Additional file 1: Table S1 for the test-retest combinations. The balanced data set made it possible to test for agreement between the pairs of BPND values only once instead of for each of the eight combinations of the test-retest and three pairs of methods of analysis (manual target ROI/automated target ROI, no realignment/realignment, manual reference ROI/automated cerebellum ROI). Hence, the statistical computations were applied once for each pair of BPND values using the mean BPND value of the other eight combinations of data (2 × 2 × 2). This averaging was justified after testing by ANOVA for the absence of two- and three-way interactions among the pairs of variables.

The two absolute variability (AV) values and ICC values were also evaluated for agreement and precision by the mean differences and confidence intervals in the same way as for the BPND values with the difference that there were 2 × 2 × 2 (i.e., eight sets or four pairs) AV and ICC values for each of the three parts of the image analysis. The comparison of ICC values between the methods was assessed by using the 95% confidence intervals [50] to examine whether there was an overlap between the confidence intervals between the pairs of methods compared.

Results

The subjects completed the study according to the protocol. Each subject underwent one T1 MRI scan on 1 day and two PET measurements in the morning and afternoon of another day after a negative urine drug screen on the same day. The time gap between PET and MRI studies was 1 to 96 days, median 6 days.

The spatial agreement of the manual and FreeSurfer ROIs estimated by the Jaccard coefficient varied between 32 and 69% (Table 1). The highest values were obtained for the caudate (59%), putamen (57%), and insula (69%), indicating that the ROIs overlapped to a major extent. The percentage of the manual ROI covered by the FreeSurfer ROI varied between 51.3 and 92.5%.

Table 1

Jaccard coefficients obtained by comparison of ROIs made manually and by FreeSurfer

ROIJaccard coefficientIntersect (voxels)Intersect % of manual volume
CAU0.5945092.5
PUT0.5761492.0
DLPFC0.32175656.2
INS0.4465279.9
CBL0.3229151.3

The ROI volumes were calculated by the number of voxels multiplied by the voxel volume for each paired ROI (both hemispheres) extracted from the ROI files. Worth noting is that the positioning of the head in the gantry differed slightly between the two PET measurements, and the subsequent two sets of re-sliced ROIs in PET space differed in volumes by 1.7 ± 3.0% (mean ± SD).

Each of the FreeSurfer ROIs and the modified FSL cerebellum ROI were significantly larger compared to the corresponding manual ROIs (Figs. 1a and ​and2,2, Table 2). The 95% CI for the differences was relatively narrow indicating a high precision for both methods (Fig. 1a, Table 2). The CoV (%) for the manual ROI volume was generally higher than for the automatically generated ROIs.

Bland-Altman plots with the mean and 95% confidence interval of ROI volumes (a) and binding potential (BPND) of [11C]SCH23390 (b) of the caudate, putamen, dorsolateral prefrontal cortex, and insula derived from manual (man) and FreeSurfer (FS) methods. Man = manual, FS = FreeSurfer, CAU = caudate, PUT = putamen, DLPFC = dorsolateral prefrontal cortex, and INS = insula

Color-coded voxels of the manual (Man) and automated (Aut) ROIs superimposed on the corresponding MR image of a healthy volunteer selected at random. ROIs generated manually are superimposed on those generated by FreeSurfer. CAU = caudate, PUT = putamen, DLPFC = dorsolateral prefrontal cortex, INS = insula, and CBL = cerebellum

Table 2

ROI volumes and binding potential (BPND) values for [11C]SCH23390 binding obtained by two repeated PET measurements analyzed by three different comparisons of methods: (1) ROI definition (manual (Man)/FreeSurfer (FS)), (2) movement correction (no realignment/realignment), and (3) definition of the cerebellum as reference region (manual/modified FSL)

CAUPUTDLPFCINSCBL
ROI volume by ROI method (mL)ManFSManFSManFSManFSManFSL
 Median6.59.58.413.039.353.710.816.67.68.1
 Interquartile range5.7–8.88.8–9.98.0–9.412.4–13.837.4–43.449.2–57.210.0–11.116.4–18.26.3–8.36.9–10.1
 CoV %16.48.912.98.415.810.09.18.019.221.1
 Mean diff (CI 95%)− 3.1 (− 3.4, − 2.8)a− 4.5 (− 5.0, −3.9)a− 12 (− 15.9, − 9.8)a− 6.7 (− 7.4, − 6.0)a− 1.0 (− 2.0, − 0.4)a
BPND by ROI methodManFSManFSManFSManFS
 Median1.651.391.851.620.320.260.550.53
 Interquartile range1.43–1.761.14–1.521.72–1.881.49–1.680.26–0.330.21–0.290.50–0.580.50–0.57
 CoV %16.120.97.08.917.019.48.88.4
 Mean diff (CI 95%)0.28 (0.24, 0.31)a0.20 (0.17, 0.23)a0.05 (0.04, 0.06)0.01 (0.01, 0.02)
BPND by realignmentNo realignRealignNo realignRealignNo realignRealignNo realignRealign
 Median1.461.571.741.730.270.300.540.53
 Interquartile range1.24–1.631.33–1.651.61–1.811.61–1.750.21–0.290.26–0.320.50–0.580.50–0.58
 CoV %22.114.79.09.721.215.48.88.2
 Mean diff (CI 95%)− 0.10 (− 0.16, − 0.05)a0.00 (− 0.03, 0.03)− 0.04 (− 0.05, − 0.02)a0.00 (− 0.01, 0.01)
BPND by reference regionMan CBLFSL CBLMan CBLFSL CBLMan CBLFSL CBLMan CBLFSL CBL
 Median1.491.531.701.770.280.290.520.55
 Interquartile range1.26–1.621.31–1.671.58–1.761.63–1.800.22–0.300.25–0.310.49–0.570.51–0.58
 CoV %18.717.68.37.318.816.98.88.3
 Mean diff (CI 95%)− 0.04 (− 0.06, − 0.03)a− 0.05 (− 0.06, − 0.03)a− 0.02 (− 0.03, − 0.01)a− 0.03 (− 0.03, − 0.01)a

The values were obtained from four brain regions in 15 subjects, each value represents the mean of the eight values for that measurement or part of the analysis. Interquartile range = 25th to 75th percent

CoV coefficient of variation, CI confidence interval

aThe confidence interval does not contain zero

As can be seen from the TACs, there was a rapid initial increase of brain radioactivity that peaked after 5–10 min followed by a rapid decline (Fig. 3). The AUC of the TACs for the realigned PET images was 2–18% higher for the manually generated target ROIs when compared to that of the FreeSurfer (p < 0.001; data not shown) (Fig. 3). The AUC of the cerebellum TACs were similar in the beginning and 3% higher during the last 24 min for the manual CBL compared to modified FSL CBL (p < 0.001 for the last 24 min; data not shown).

Time-activity curves from a healthy volunteer after i.v. injection of [11C]SCH23390 in the morning (left) and afternoon (right). The curves were derived from realigned dynamic PET images using manually and automatically generated ROIs, respectively. PUT = putamen, DLPFC = dorsolateral prefrontal cortex, CBL = cerebellum. SUV = standardized uptake value = (CROI/inj dose) × body weight

The eight sets of target TACs and two sets of cerebellum TACs amounted to 16 BPND values for each ROI (Additional file 1: Table S1). The statistical computations of the regional BPND values for the three parts of the analysis are listed in Table 2. The 95% CI for most of the differences was relatively narrow and did not contain zero indicating statistical significance for the difference between methods (Fig. 1b). The manually generated BPND values were 3–21% higher with a generally low CoV and interquartile range compared to that of the automatically generated BPND values. After realignment, the BPND values became higher for the caudate and DLPFC (Table 2). On the contrary, the BPND values for putamen and insula did not increase after realignment. The use of the modified FSL cerebellum ROI increased BPND by 3 to 7% in all regions as compared to that of the manual CBL.

The absolute variability values (AV) for each of the four regional BPND values were calculated for the total of eight pairs of regional BPND values from the three parts of the analysis (2 × 2 × 2 = 8) (Additional file 1: Table S1). The AV values among the methods varied between 3.9–11.9% for the caudate and putamen and 8.0–17.9% for the DLPFC and insula. The 95% CI of the mean difference in AV was wide and included zero in most cases indicating no significant differences (Table 3). However, following realignment, the 95% CI of the mean difference in AV did not overlap with zero for caudate, putamen, and DLPFC, suggesting a small but significant improvement. The use of automated cerebellum did not improve the AV values in any region compared to that of the manual. The ICCs were generally higher using FreeSurfer, but all comparisons fell within the limits of the 95% confidence intervals, meaning that these differences are not significant.

Table 3

Two sets of mean absolute variability (AV) values and intraclass correlation coefficients (ICC) of the binding potential (BPND) of [11C]SCH23390 in four brain regions derived from two different methods (each defined in the leftmost column) in each of three parts of the image analysis process: (1) manual/FreeSurfer ROIs, (2) no realignment/realignment, and (3) manual cerebellum ROI/automated FSL cerebellum ROI)

Method (1/2)ROIAV (%) method 1AV (%) method 2AV diff (95% CI)ICC (95% CI) Method 1ICC (95% CI) Method 2
ROI definition
Method 1: manual
Method 2: FreeSurfer		CAU9.459.791.61 (− 1.96, 1.27)0.79 (0.49, 0.92)0.87 (0.66, 0.95)
PUT4.824.720.95 (− 0.86, 1.05)0.73 (0.38, 0.90)0.84 (0.59, 0.94)
DLPFC13.1714.302.67 (− 3.79, 1.55)0.72 (0.36, 0.89)0.72 (0.36, 0.90)
INS9.737.941.76 (0.03, 3.54)0.31 (− 0.25, 0.70)0.46 (− 0.07, 0.78)
Movement correction
Method 1: none
Method 2: realignment		CAU13.786.814.43 (2.55, 11.40)0.79 (0.50, 0.93)0.86 (0.64, 0.95)
PUT5.883.761.68 (0.45, 3.80)0.75 (0.40, 0.91)0.81 (0.53, 0.93)
DLPFC19.4810.086.08 (3.32, 15.48)0.63 (0.21, 0.86)0.77 (0.44, 0.91)
INS9.408.471.50 (− 0.56, 2.44)0.38 (− 0.18, 0.74)0.37 (− 0.18, 0.74)
Definition of reference region (cerebellum)
Method 1: manual
Method 2: FSL		CAU10.609.151.86 (− 0.41, 3.30)0.81 (0.53, 0.93)0.84 (0.59, 0.94)
PUT5.374.382.04 (1.05, 3.03)0.77 (0.45, 0.91)0.76 (0.41, 0.91)
DLPFC15.2911.546.08 (− 2.33, 9.82)0.71 (0.33, 0.89)0.73 (0.38, 0.90)
INS10.737.882.81 (0.05, 5.66)0.23 (−0.34, 0.66)0.50 (− 0.02, 0.80)

Discussion

The main objective of the present study was to examine the impact of specific methodological steps used in quantitative image analysis of [11C]SCH23390 binding to D1-DR on variability and reliability. The methods were chosen to allow for comparison to historical conditions so that the results would be applicable for interpreting the existing PET literature. For that purpose, data were acquired using the PET system ECAT EXACT HR, which has a resolution comparable to the currently widely used whole-body PET systems. In relation to the major aims of the study, the results show that automated ROIs generally produce lower BPND values than manual ROIs, whereas absolute variability was similar. Moreover, the use of realignment improved the absolute variability, and an automated cerebellum ROI yielded slightly increased BPND values.

The volumes were significantly larger for the automated ROIs, with lower interindividual variability except for the cerebellum, which had larger variability for the automated ROI (Table 2). The differences in volumes between the methods were largest for the caudate and putamen. Similar results of smaller manual compared to FreeSurfer striatal ROIs, but with greater interindividual variability, have recently been reported when analyzing MRI data for the caudate and putamen [, ]. Other studies have demonstrated that FS can produce measurements that are comparable to those derived from manual tracing of ROIs [, ]. However, manual editing of FS ROIs may still be required in order to improve validity [].

The intersection of the manual and FreeSurfer ROIs was smaller than the manual ROI itself, which indicates that the ROIs did not overlap completely or in a symmetrical fashion (Table 1). Hence, to shrink or erode the FreeSurfer ROI, as shown for partial volume correction [], to the same volume as the manual would still result in a placement mismatch.

FreeSurfer has become the standard for obtaining cortical metrics with demonstrated high validity and reliability []. For the subcortical regions, manual segmentation still remains the gold standard due to a better specificity of anatomic boundaries compared to FreeSurfer which yields larger caudate and putamen and to FSL-FIRST generating larger caudate ROIs [, ]. However, manual subcortical ROI volumes do also vary between different raters []. Hence, the differences in the size and placement of the ROIs in the present study may be due to both systematic differences in anatomical designation and in software-based segmentation.

The higher BPND for the manual ROIs is consistent with their smaller volumes (Table 2). All regions examined have a higher BP when compared with surrounding tissue or the subarachnoidal space. More spill-out than spill-in of the measured radioactivity will yield a recovery coefficient < 1. It is thus conceivable that recovery and BPND will be lower for a larger delineation of a particular ROI.

After realignment of the dynamic PET images, the BPND values were significantly higher for the caudate and the dorsolateral prefrontal cortex. For putamen and insula, there was no evident effect of realignment on the BPND values. A possible explanation for the regional differences in the impact of realignment is that the caudate and dorsolateral prefrontal cortex border to the subarachnoidal or ventricular space where radioactivity is very low or negligible. These regions are thus more sensitive to movement artifacts, while the putamen and insula are embedded in the white matter and subsequently less sensitive.

The use of the modified FSL ROIs for CBL increased the BPND values by a few percents and decreased the corresponding CoVs for all regions compared to those obtained using the manual cerebellum (Table 2). A possible explanation is that the trimmed and masked FSL ROIs for CBL are less sensitive to the partial volume effects. However, the small difference and narrow CI indicate similarity between these measures and hence the usefulness of both the manual and automated cerebellum ROI.

Generally, the absolute variability decreased significantly after realignment, likely due to the influence of more reliable TACs on the subsequent calculation of the BPND. For the CAU and PUT, which were the regions showing the highest degree of anatomical overlap in our study and therefore represents the best points of comparison, the absolute variability was 6.47 and 3.54% respectively for the automated ROIs (Additional file 1: Table S1). These observations are in line with two previous test-retest studies using manual ROIs to quantify [11C]SCH23390 binding. Hirvonen and coworkers investigated five healthy volunteers and found an absolute variability of 4.2–6.6% [], whereas the absolute variability was 7.8–8.4% in a recent study on 13 individuals []. Importantly, the present study shows that the use of automated ROIs has a similar absolute variability of D1-DR BPND values as compared to that of manual ROIs.

The present level of absolute variability among all the brain regions is comparable to that of other PET neuroreceptor radioligands such as [11C]raclopride [, ], [11C]FLB 457 [], [11C]MADAM [], and [11C]AZ10419369 []. The repeatability of PET-measured BPND is of importance to increase the statistical power in receptor occupancy studies, as well as in both cross-sectional and longitudinal studies. For instance, a power analysis based on the present results gives a statistical power to detect a 20% change in BPND (n = 10) in the putamen of 100 and 96% for the manual and FreeSurfer methods, respectively. A 20% change represents a Cohen’s d of over 6, which is extremely high for biological changes but commonly seen in drug receptor occupancy studies.

Whereas significant ventricular enlargement and cerebral cortical atrophy are a well-replicated observation in schizophrenia, these changes are usually small and require large sample sizes for detection [, ]. Hence, the present findings in healthy volunteers should be applicable in schizophrenia research. Importantly, our results do not support a role for different ROI delineation methods as an explanation for discrepant results in studies on D1-R in drug-naïve schizophrenia. In contrast, the present increase in BPND in putamen and DLPFC with improved absolute variability after realignment may be more relevant. In studies in drug-naïve samples, Okubo et al. found lower D1-R in the frontal cortex in patients, whereas Abi-Dargham and colleagues found higher levels [, ] and Karlsson et al. found no significant differences, although D1-R was numerically higher with an effect size of 0.3 in the frontal cortex [, ]. Across all of these papers, movement correction is only described in the 2002 study by Abi-Dargham et al. [], but no information is provided in the 2012 paper []. Although both our previous and the present studies employed head fixation to minimize head movement, the present results show that the absence of movement correction may still produce both lower BPND and lower reliability. It should be noted that the magnitude of change in BPND values following movement correction would not be sufficiently large to explain the differences in the direction of the reported clinical findings. However, it may be speculated that movement artifacts may have occurred more commonly among patients than healthy volunteers and subsequently biased the results.

In contrast to the subtle structural brain changes in schizophrenia, gross changes are consistently seen in neurodegenerative diseases such as Alzheimer’s, Parkinson’s, and Huntington’s diseases [–]. Structural brain changes may be biased differently by the manual and automated methods, as shown in a recent cross-validation study among automated and manual ROI methods on patients with multiple sclerosis where differences in atrophy measurements both in size and proportionality were demonstrated []. Although FreeSurfer has functionality for manual intervention to accommodate for certain morphological changes [], the present results of similar absolute variability of manual and automated FresSurfer ROI definition (for PET) in healthy subjects need to be confirmed in patient with neurodegenerative disorders.

There are two major advantages of automated ROI definition. First, it is less labor-intensive than tedious hand tracing of ROIs, especially in large high-resolution data sets, and second, it is less biased to rater subjectivity when compared to manual methods. Additionally, automated ROI definition is more amenable for pooling data for meta-analysis, an important aspect since PET studies in patient populations usually have small sample sizes.

Conclusions

In summary, the results show that the repeatability of BPND was similar between the manual and FS ROIs, while it improved after realignment in all regions. The results suggest that the choice of the ROI method is more dependent on questions of validity, such as anatomical precision, which may be particularly important in patients with gross morphological changes or very localized pathology. The importance of the correction of motion artifacts by realignment to obtain higher and more reliable BPND values may be particularly important in patient populations such as schizophrenia where motion artifacts may be more prevalent.

Additional file

Additional file 1:(19K, docx)

Table S1. Mean BPND values of [11C]SCH23390 in the morning and afternoon, the absolute variability, and ICC in 15 healthy men in four brain regions. There are eight BPND values for each region derived from the different combinations of methods in the analysis process. (DOCX 18 kb)

Acknowledgements

The authors thank the Karolinska positron emission tomography group for the excellent technical assistance and in particular Dr. Zhisheng Jia and Guennadi Jogolev.

Funding

This study was funded by the Swedish Science Council grant no. 2015-02398 to LF and by Pfizer, Inc. The funding bodies did not play a role in the design of the study; collection, analysis, and interpretation of data; and in writing the manuscript.

Availability of data and materials

Please contact the author for data requests.

Authors’ contributions

LF, CH, and PS conceived the study. All authors designed the study. PS did the screening and examination of the subjects. LF, GJM, and PS analyzed the data and interpreted the results. LF and PS drafted the article. All authors critically revised the article and approved of the final version for publication.

Notes

Ethics approval and consent to participate

The study was conducted in compliance with the ethical principles originating in the Declaration of Helsinki and the Good Clinical Practice (GCP) Guidelines of the International Conference on Harmonization (ICH). The study was approved by the Regional Ethical Review Board in Stockholm, Sweden (registration number 2013/1703-31/1) and the Radiation Safety Committee at the Karolinska University Hospital in Solna, Stockholm, Sweden (registration number K1865-2013). Written informed consent to participate was obtained from all individual participants.

Competing interests

LF is a shareholder of AstraZeneca. All other authors declare that they have no competing interest.

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Contributor Information

Per Stenkrona, Phone: +46 8 51772902, Email: es.ik@anorknets.rep.

Granville J. Matheson, Email: es.ik@nosehtam.ellivnarg.

Simon Cervenka, Email: es.ik@aknevrec.nomis.

Pontus Plavén Sigray, Email: es.ik@yargis-nevalp.sutnop.

Christer Halldin, Email: es.ik@nidllah.retsirhc.

Lars Farde, Email: es.ik@edraf.sral.

References

1. Wagner HN, Jr, Burns HD, Dannals RF, Wong DF, Langstrom B, Duelfer T, et al. Imaging dopamine receptors in the human brain by positron tomography. Science. 1983;221:1264–1266. doi: 10.1126/science.6604315. [PubMed] [CrossRef] [Google Scholar]
2. Garnett ES, Firnau G, Nahmias C. Dopamine visualized in the basal ganglia of living man. Nature. 1983;305:137–138. doi: 10.1038/305137a0. [PubMed] [CrossRef] [Google Scholar]
3. Howes O, McCutcheon R, Stone J. Glutamate and dopamine in schizophrenia: an update for the 21st century. J Psychopharmacol. 2015;29:97–115. doi: 10.1177/0269881114563634.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
4. Kambeitz J, Abi-Dargham A, Kapur S, Howes OD. Alterations in cortical and extrastriatal subcortical dopamine function in schizophrenia: systematic review and meta-analysis of imaging studies. Br J Psychiatry. 2014;204:420–429. doi: 10.1192/bjp.bp.113.132308. [PubMed] [CrossRef] [Google Scholar]
5. Abi-Dargham A, Mawlawi O, Lombardo I, Gil R, Martinez D, Huang Y, et al. Prefrontal dopamine D1 receptors and working memory in schizophrenia. J Neurosci. 2002;22:3708–3719. doi: 10.1523/JNEUROSCI.22-09-03708.2002. [PubMed] [CrossRef] [Google Scholar]
6. Abi-Dargham A, Xu X, Thompson JL, Gil R, Kegeles LS, Urban N, et al. Increased prefrontal cortical D(1) receptors in drug naive patients with schizophrenia: a PET study with [(1)(1)C]NNC112. J Psychopharmacol. 2012;26:794–805. doi: 10.1177/0269881111409265. [PubMed] [CrossRef] [Google Scholar]
7. Okubo Y, Suhara T, Suzuki K, Kobayashi K, Inoue O, Terasaki O, et al. Decreased prefrontal dopamine D1 receptors in schizophrenia revealed by PET. Nature. 1997;385:634–636. doi: 10.1038/385634a0. [PubMed] [CrossRef] [Google Scholar]
8. Karlsson P, Farde L, Halldin C, Sedvall G. PET study of D(1) dopamine receptor binding in neuroleptic-naive patients with schizophrenia. Am J Psychiatry. 2002;159:761–767. doi: 10.1176/appi.ajp.159.5.761. [PubMed] [CrossRef] [Google Scholar]
9. Kosaka J, Takahashi H, Ito H, Takano A, Fujimura Y, Matsumoto R, et al. Decreased binding of [11C]NNC112 and [11C]SCH23390 in patients with chronic schizophrenia. Life Sci. 2010;86:814–818. doi: 10.1016/j.lfs.2010.03.018. [PubMed] [CrossRef] [Google Scholar]
10. Hirvonen J, van Erp TG, Huttunen J, Aalto S, Nagren K, Huttunen M, et al. Brain dopamine d1 receptors in twins discordant for schizophrenia. Am J Psychiatry. 2006;163:1747–1753. doi: 10.1176/ajp.2006.163.10.1747. [PubMed] [CrossRef] [Google Scholar]
11. Lidow MS, Goldman-Rakic PS. A common action of clozapine, haloperidol, and remoxipride on D1- and D2-dopaminergic receptors in the primate cerebral cortex. Proc Natl Acad Sci U S A. 1994;91:4353–4356. doi: 10.1073/pnas.91.10.4353.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
12. Lidow MS, Elsworth JD, Goldman-Rakic PS. Down-regulation of the D1 and D5 dopamine receptors in the primate prefrontal cortex by chronic treatment with antipsychotic drugs. J Pharmacol Exp Ther. 1997;281:597–603. [PubMed] [Google Scholar]
13. Cervenka S. PET radioligands for the dopamine D1-receptor: application in psychiatric disorders. Neurosci Lett. 2018; 10.1016/j.neulet.2018.03.007. [PubMed]
14. Poels EM, Girgis RR, Thompson JL, Slifstein M, Abi-Dargham A. In vivo binding of the dopamine-1 receptor PET tracers [(1)(1)C]NNC112 and [(1)(1)C]SCH23390: a comparison study in individuals with schizophrenia. Psychopharmacology. 2013;228:167–174. doi: 10.1007/s00213-013-3026-8. [PubMed] [CrossRef] [Google Scholar]
15. Klein A, Andersson J, Ardekani BA, Ashburner J, Avants B, Chiang MC, et al. Evaluation of 14 nonlinear deformation algorithms applied to human brain MRI registration. NeuroImage. 2009;46:786–802. doi: 10.1016/j.neuroimage.2008.12.037.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
16. Chow TW, Takeshita S, Honjo K, Pataky CE, St Jacques PL, Kusano ML, et al. Comparison of manual and semi-automated delineation of regions of interest for radioligand PET imaging analysis. BMC Nucl Med. 2007;7:2. doi: 10.1186/1471-2385-7-2.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
17. Raniga P, Bourgeat P, Fripp J, Acosta O, Villemagne VL, Rowe C, et al. Automated (11)C-PiB standardized uptake value ratio. Acad Radiol. 2008;15:1376–1389. doi: 10.1016/j.acra.2008.07.006. [PubMed] [CrossRef] [Google Scholar]
18. Rosario BL, Weissfeld LA, Laymon CM, Mathis CA, Klunk WE, Berginc MD, et al. Inter-rater reliability of manual and automated region-of-interest delineation for PiB PET. NeuroImage. 2011;55:933–941. doi: 10.1016/j.neuroimage.2010.12.070.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
19. Su Y, D’Angelo GM, Vlassenko AG, Zhou G, Snyder AZ, Marcus DS, et al. Quantitative analysis of PiB-PET with FreeSurfer ROIs. PLoS One. 2013;8:e73377. doi: 10.1371/journal.pone.0073377.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
20. Wallius E, Tohka J, Hirvonen J, Hietala J, Ruotsalainen U. Evaluation of the automatic three-dimensional delineation of caudate and putamen for PET receptor occupancy studies. Nucl Med Commun. 2008;29:53–65. doi: 10.1097/MNM.0b013e3282f1bba0. [PubMed] [CrossRef] [Google Scholar]
21. Johansson J, Alakurtti K, Joutsa J, Tohka J, Ruotsalainen U, Rinne JO. Comparison of manual and automatic techniques for substriatal segmentation in 11C-raclopride high-resolution PET studies. Nucl Med Commun. 2016;37:1074–1087. doi: 10.1097/MNM.0000000000000559. [PubMed] [CrossRef] [Google Scholar]
22. Johansson J, Alakurtti K, Joutsa J, Tohka J, Ruotsalainen U, Rinne JO. Comparison of manual and automatic techniques for substriatal segmentation in 11C-raclopride high-resolution PET studies. Nucl Med Commun. 2016; 10.1097/MNM.0000000000000559. [PubMed]
23. Wienhard K, Dahlbom M, Eriksson L, Michel C, Bruckbauer T, Pietrzyk U, et al. The ECAT EXACT HR: performance of a new high resolution positron scanner. J Comput Assist Tomogr. 1994;18:110–118. doi: 10.1097/00004728-199401000-00023. [PubMed] [CrossRef] [Google Scholar]
24. Bergstrom M, Boethius J, Eriksson L, Greitz T, Ribbe T, Widen L. Head fixation device for reproducible position alignment in transmission CT and positron emission tomography. J Comput Assist Tomogr. 1981;5:136–141. doi: 10.1097/00004728-198102000-00027. [PubMed] [CrossRef] [Google Scholar]
25. Halldin C, Stone-Elander S, Farde L, Ehrin E, Fasth KJ, Langstrom B, et al. Preparation of 11C-labelled SCH 23390 for the in vivo study of dopamine D-1 receptors using positron emission tomography. Int J Radiat Appl Instrum Part A. 1986;37:1039–1043. doi: 10.1016/0883-2889(86)90044-4. [PubMed] [CrossRef] [Google Scholar]
26. Schain M, Toth M, Cselenyi Z, Stenkrona P, Halldin C, Farde L, et al. Quantification of serotonin transporter availability with [11C]MADAM—a comparison between the ECAT HRRT and HR systems. NeuroImage. 2012;60:800–807. doi: 10.1016/j.neuroimage.2011.12.047. [PubMed] [CrossRef] [Google Scholar]
27. Friston KJ, Ashburner J, Frith CD, Poline J-B, Heather JD, Frackowiak RSJ. Spatial registration and normalization of images. Hum Brain Mapp. 1995;3:165–189. doi: 10.1002/hbm.460030303. [CrossRef] [Google Scholar]
28. Studholme C, Hill D, Hawkes DJ. Normalized entropy measure of 3-D medical image alignment. Medical imaging 1998. San Diego; 1998. p. 132–43.
29. Roland PE, Graufelds CJ, Wåhlin L, Ingelman L, Andersson M, Ledberg A, et al. Human brain atlas: for high-resolution functionel and anatomical mapping. Hum Brain Mapp. 1994;1:173–184. doi: 10.1002/hbm.460010303. [PubMed] [CrossRef] [Google Scholar]
30. Mawlawi O, Martinez D, Slifstein M, Broft A, Chatterjee R, Hwang DR, et al. Imaging human mesolimbic dopamine transmission with positron emission tomography: I. Accuracy and precision of D(2) receptor parameter measurements in ventral striatum. J Cereb Blood Flow Metab. 2001;21:1034–1057. doi: 10.1097/00004647-200109000-00002. [PubMed] [CrossRef] [Google Scholar]
31. Fischl B. FreeSurfer. NeuroImage. 2012;62:774–781. doi: 10.1016/j.neuroimage.2012.01.021.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
32. Dale AM, Fischl B, Sereno MI. Cortical surface-based analysis. I Segmentation Surface Reconstruction Neuroimage. 1999;9:179–194. [PubMed] [Google Scholar]
33. Fischl B, Sereno MI, Dale AM. Cortical surface-based analysis. II: inflation, flattening, and a surface-based coordinate system. NeuroImage. 1999;9:195–207. doi: 10.1006/nimg.1998.0396. [PubMed] [CrossRef] [Google Scholar]
34. Han X, Jovicich J, Salat D, van der Kouwe A, Quinn B, Czanner S, et al. Reliability of MRI-derived measurements of human cerebral cortical thickness: the effects of field strength, scanner upgrade and manufacturer. NeuroImage. 2006;32:180–194. doi: 10.1016/j.neuroimage.2006.02.051. [PubMed] [CrossRef] [Google Scholar]
35. Reuter M, Schmansky NJ, Rosas HD, Fischl B. Within-subject template estimation for unbiased longitudinal image analysis. NeuroImage. 2012;61:1402–1418. doi: 10.1016/j.neuroimage.2012.02.084.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
36. Rosas HD, Liu AK, Hersch S, Glessner M, Ferrante RJ, Salat DH, et al. Regional and progressive thinning of the cortical ribbon in Huntington’s disease. Neurology. 2002;58:695–701. doi: 10.1212/WNL.58.5.695. [PubMed] [CrossRef] [Google Scholar]
37. Kuperberg GR, Broome MR, McGuire PK, David AS, Eddy M, Ozawa F, et al. Regionally localized thinning of the cerebral cortex in schizophrenia. Arch Gen Psychiatry. 2003;60:878–888. doi: 10.1001/archpsyc.60.9.878. [PubMed] [CrossRef] [Google Scholar]
38. Fischl B, Sereno MI, Tootell RB, Dale AM. High-resolution intersubject averaging and a coordinate system for the cortical surface. Hum Brain Mapp. 1999;8:272–284. doi: 10.1002/(SICI)1097-0193(1999)8:4<272::AID-HBM10>3.0.CO;2-4. [PubMed] [CrossRef] [3.0.CO;2-4&' target='pmc_ext' ref='reftype=other&article-id=6070454&issue-id=303786&journal-id=1625&FROM=Article%7CCitationRef&TO=Content%20Provider%7CLink%7CGoogle%20Scholar'>Google Scholar]
39. Desikan RS, Segonne F, Fischl B, Quinn BT, Dickerson BC, Blacker D, et al. An automated labeling system for subdividing the human cerebral cortex on MRI scans into gyral based regions of interest. NeuroImage. 2006;31:968–980. doi: 10.1016/j.neuroimage.2006.01.021. [PubMed] [CrossRef] [Google Scholar]
40. Jenkinson M, Beckmann CF, Behrens TE, Woolrich MW, Smith SM. Fsl. NeuroImage. 2012;62:782–790. doi: 10.1016/j.neuroimage.2011.09.015. [PubMed] [CrossRef] [Google Scholar]
41. Diedrichsen J, Balsters JH, Flavell J, Cussans E, Ramnani N. A probabilistic MR atlas of the human cerebellum. NeuroImage. 2009;46:39–46. doi: 10.1016/j.neuroimage.2009.01.045. [PubMed] [CrossRef] [Google Scholar]
42. Matheson GJ, Stenkrona P, Cselenyi Z, Plaven-Sigray P, Halldin C, Farde L, et al. Reliability of volumetric and surface-based normalisation and smoothing techniques for PET analysis of the cortex: a test-retest analysis using [11C]SCH-23390. NeuroImage. 2017;155:344–353. doi: 10.1016/j.neuroimage.2017.04.031. [PubMed] [CrossRef] [Google Scholar]
43. Lammertsma AA, Hume SP. Simplified reference tissue model for PET receptor studies. NeuroImage. 1996;4:153–158. doi: 10.1006/nimg.1996.0066. [PubMed] [CrossRef] [Google Scholar]
44. Kessler LG, Barnhart HX, Buckler AJ, Choudhury KR, Kondratovich MV, Toledano A, et al. The emerging science of quantitative imaging biomarkers terminology and definitions for scientific studies and regulatory submissions. Stat Methods Med Res. 2015;24:9–26. doi: 10.1177/0962280214537333. [PubMed] [CrossRef] [Google Scholar]
45. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet. 1986;1:307–310. doi: 10.1016/S0140-6736(86)90837-8. [PubMed] [CrossRef] [Google Scholar]
46. du Prel JB, Hommel G, Rohrig B, Blettner M. Confidence interval or p-value?: part 4 of a series on evaluation of scientific publications. Dtsch Arztebl Int. 2009;106:335–339.[PMC free article] [PubMed] [Google Scholar]
47. Obuchowski NA, Reeves AP, Huang EP, Wang XF, Buckler AJ, Kim HJ, et al. Quantitative imaging biomarkers: a review of statistical methods for computer algorithm comparisons. Stat Methods Med Res. 2015;24:68–106. doi: 10.1177/0962280214537390.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
48. Thirthalli J, Rajkumar RP. Statistical versus clinical significance in psychiatric research—an overview for beginners. Asian J Psychiatr. 2009;2:74–79. doi: 10.1016/j.ajp.2009.04.004. [PubMed] [CrossRef] [Google Scholar]
49. Shakespeare TP, Gebski VJ, Veness MJ, Simes J. Improving interpretation of clinical studies by use of confidence levels, clinical significance curves, and risk-benefit contours. Lancet. 2001;357:1349–1353. doi: 10.1016/S0140-6736(00)04522-0. [PubMed] [CrossRef] [Google Scholar]
50. Searle SR. Linear models. New York: Wiley; 1971. [Google Scholar]
51. Perlaki G, Horvath R, Nagy SA, Bogner P, Doczi T, Janszky J, et al. Comparison of accuracy between FSL’s FIRST and FreeSurfer for caudate nucleus and putamen segmentation. Sci Rep. 2017;7:2418. doi: 10.1038/s41598-017-02584-5.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
52. Fischl B, Salat DH, Busa E, Albert M, Dieterich M, Haselgrove C, et al. Whole brain segmentation: automated labeling of neuroanatomical structures in the human brain. Neuron. 2002;33:341–355. doi: 10.1016/S0896-6273(02)00569-X. [PubMed] [CrossRef] [Google Scholar]
53. Tae WS, Kim SS, Lee KU, Nam EC, Kim KW. Validation of hippocampal volumes measured using a manual method and two automated methods (FreeSurfer and IBASPM) in chronic major depressive disorder. Neuroradiology. 2008;50:569–581. doi: 10.1007/s00234-008-0383-9. [PubMed] [CrossRef] [Google Scholar]
54. McCarthy CS, Ramprashad A, Thompson C, Botti JA, Coman IL, Kates WR. A comparison of FreeSurfer-generated data with and without manual intervention. Front Neurosci. 2015;9:379. doi: 10.3389/fnins.2015.00379.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
55. Rousset O, Rahmim A, Alavi A, Zaidi H. Partial volume correction strategies in PET. PET Clin. 2007;2:235–249. doi: 10.1016/j.cpet.2007.10.005. [PubMed] [CrossRef] [Google Scholar]
56. Cardinale F, Chinnici G, Bramerio M, Mai R, Sartori I, Cossu M, et al. Validation of FreeSurfer-estimated brain cortical thickness: comparison with histologic measurements. Neuroinformatics. 2014;12:535–542. doi: 10.1007/s12021-014-9229-2. [PubMed] [CrossRef] [Google Scholar]
57. Nugent AC, Luckenbaugh DA, Wood SE, Bogers W, Zarate CA, Jr, Drevets WC. Automated subcortical segmentation using FIRST: test-retest reliability, interscanner reliability, and comparison to manual segmentation. Hum Brain Mapp. 2013;34:2313–2329. doi: 10.1002/hbm.22068.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
58. Hirvonen J, Nagren K, Kajander J, Hietala J. Measurement of cortical dopamine d1 receptor binding with 11C[SCH23390]: a test-retest analysis. J Cereb Blood Flow Metab. 2001;21:1146–1150. doi: 10.1097/00004647-200110000-00002. [PubMed] [CrossRef] [Google Scholar]
59. Kaller S, Rullmann M, Patt M, Becker GA, Luthardt J, Girbardt J, et al. Test-retest measurements of dopamine D1-type receptors using simultaneous PET/MRI imaging. Eur J Nucl Med Mol Imaging. 2017;44:1025–1032. doi: 10.1007/s00259-017-3645-0. [PubMed] [CrossRef] [Google Scholar]
60. Nyberg S, Farde L, Halldin C. Test-retest reliability of central [11C]raclopride binding at high D2 receptor occupancy. A PET study in haloperidol-treated patients. Psychiatry Res. 1996;67:163–171. doi: 10.1016/0925-4927(96)02921-6. [PubMed] [CrossRef] [Google Scholar]
61. Sudo Y, Suhara T, Inoue M, Ito H, Suzuki K, Saijo T, et al. Reproducibility of [11 C]FLB 457 binding in extrastriatal regions. Nucl Med Commun. 2001;22:1215–1221. doi: 10.1097/00006231-200111000-00008. [PubMed] [CrossRef] [Google Scholar]
62. Lundberg J, Halldin C, Farde L. Measurement of serotonin transporter binding with PET and [11C]MADAM: a test-retest reproducibility study. Synapse. 2006;60:256–263. doi: 10.1002/syn.20297. [PubMed] [CrossRef] [Google Scholar]
63. Nord M, Finnema SJ, Schain M, Halldin C, Farde L. Test-retest reliability of [11C]AZ10419369 binding to 5-HT(1B) receptors in human brain. Eur J Nucl Med Mol Imaging. 2014;41:301–307. doi: 10.1007/s00259-013-2529-1. [PubMed] [CrossRef] [Google Scholar]
64. Haukvik UK, Hartberg CB, Agartz I. Schizophrenia—what does structural MRI show? Tidsskr Nor Laegeforen. 2013;133:850–853. doi: 10.4045/tidsskr.12.1084. [PubMed] [CrossRef] [Google Scholar]
65. Olabi B, Ellison-Wright I, McIntosh AM, Wood SJ, Bullmore E, Lawrie SM. Are there progressive brain changes in schizophrenia? A meta-analysis of structural magnetic resonance imaging studies. Biol Psychiatry. 2011;70:88–96. doi: 10.1016/j.biopsych.2011.01.032. [PubMed] [CrossRef] [Google Scholar]
66. Teipel SJ, Grothe M, Lista S, Toschi N, Garaci FG, Hampel H. Relevance of magnetic resonance imaging for early detection and diagnosis of Alzheimer disease. Med Clin North Am. 2013;97:399–424. doi: 10.1016/j.mcna.2012.12.013. [PubMed] [CrossRef] [Google Scholar]
67. Politis M. Neuroimaging in Parkinson disease: from research setting to clinical practice. Nat Rev Neurol. 2014;10:708–722. doi: 10.1038/nrneurol.2014.205. [PubMed] [CrossRef] [Google Scholar]
68. Aylward EH. Magnetic resonance imaging striatal volumes: a biomarker for clinical trials in Huntington’s disease. Mov Disord. 2014;29:1429–1433. doi: 10.1002/mds.26013.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
69. Steenwijk MD, Amiri H, Schoonheim MM, de Sitter A, Barkhof F, Pouwels PJW, et al. Agreement of MSmetrix with established methods for measuring cross-sectional and longitudinal brain atrophy. Neuroimage Clin. 2017;15:843–853. doi: 10.1016/j.nicl.2017.06.034.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
Articles from EJNMMI Research are provided here courtesy of Springer-Verlag
Published online 2015 Aug 14. doi: 10.1007/s13311-015-0374-6
PMID: 26271952
This article has been cited by other articles in PMC.

Associated Data

Supplementary Materials
13311_2015_374_MOESM1_ESM.pdf (221K)
13311_2015_374_MOESM2_ESM.pdf (221K)

Abstract

The endocannabinoid system is currently defined as the ensemble of the two 7-transmembrane-domain and G protein-coupled receptors for Δ9-tetrahydrocannabinol (but not for most other plant cannabinoids or phytocannabinoids)—cannabinoid receptor type-1 (CB1R) and cannabinoid receptor type-2 (CB2R); their two most studied endogenous ligands, the “endocannabinoids” N-arachidonoylethanolamine (anandamide) and 2-arachidonoylglycerol (2-AG); and the enzymes responsible for endocannabinoid metabolism. However, anandamide and 2-AG, and also the phytocannabinoids, have more molecular targets than just CB1R and CB2R. Furthermore, the endocannabinoids, like most other lipid mediators, have more than just one set of biosynthetic and degrading pathways and enzymes, which they often share with “endocannabinoid-like” mediators that may or may not interact with the same proteins as Δ9-tetrahydrocannabinol and other phytocannabinoids. In some cases, these degrading pathways and enzymes lead to molecules that are not inactive and instead interact with other receptors. Finally, some of the metabolic enzymes may also participate in the chemical modification of molecules that have very little to do with endocannabinoid and cannabinoid targets. Here, we review the whole world of ligands, receptors, and enzymes, a true “endocannabinoidome”, discovered after the cloning of CB1R and CB2R and the identification of anandamide and 2-AG, and its interactions with phytocannabinoids.

Electronic supplementary material

The online version of this article (doi:10.1007/s13311-015-0374-6) contains supplementary material, which is available to authorized users.

Keywords: Phytocannabinoids, Endocannabinoids, TRP channels, Endocannabinoidome

Introduction

Giving a definition of the complex endogenous signaling system known as the “endocannabinoid system” is becoming an increasingly difficult task. In fact, the number of potential components of this system, which was originally identified from studies on the mechanism of action of the psychotropic ingredient of some varieties of cannabis, Δ9-tetrahydrocannabinol (THC), is increasing with the passing years, and the definition of “endocannabinoid” is also bound to change in the near future [1].

At the turn of the century, the endocannabinoid system was defined as the ensemble of 1) two 7-transmembrane-domain and G protein-coupled receptors (GPCRs) for THC—cannabinoid receptor type-1 (CB1R) and cannabinoid receptor type-2 (CB2R); 2) their 2 most studied endogenous ligands, the “endocannabinoids” N-arachidonoylethanolamine (anandamide) and 2-arachidonoylglycerol (2-AG); and 3) the 5 enzymes believed, at that time, to be uniquely responsible for endocannabinoid biosynthesis [i.e., N-acyl-phosphatidyl-ethanolamine-selective phospholipase D (NAPE-PLD) and diacylglycerol lipases (DAGL) α and β, for anandamide and 2-AG, respectively] and hydrolytic inactivation [i.e., fatty acid amide hydrolase (FAAH) and monoacylglycerol lipase (MAGL), for anandamide and 2-AG, respectively] (Fig. 1) []. This definition, however, presented a few semantic problems [1]: 1) of the > 80 cannabinoids naturally found in cannabis (with different relative composition depending on the cannabis variety), only THC and its less abundant propyl analogue, Δ9-tetrahydrocannabivarin (THCV), are capable of binding with high affinity to CB1R and CB2R (with agonist and antagonist activity for THC and THCV, respectively); hence, these 2 receptors should not be defined as “cannabinoid” receptors, but rather as THC/THCV receptors [alternatively the definition of “cannabinoid receptor” should also include those proteins that often bind to cannabinoids, such as the thermosensitive transient receptor potential (TRP) cation channels (thermo-TRPs) [] (see below)]; 2) as a consequence, “endocannabinoids” should not be the endogenous ligands of CB1R and CB2R, but rather the ligands of all those “cannabinoid receptors” that uniquely and selectively bind to cannabinoids in general (thus, anandamide and 2-AG might not be the only endocannabinoids); and 3) again, as a consequence, “endocannabinoid enzymes” would not only be NAPE-PLD, the two DAGLs, FAAH, and MAGL, but also other enzymes responsible for the biosynthesis and inactivation of the other mediators to be eventually included in the list of the endocannabinoids.

Complexity, redundancy, and promiscuity of the endocannabinoid system: the “endocannabinodome” and the interactions therewith of plant cannabinoids. Several often concurrent pathways underlie both the biosynthesis and the inactivation of the 2 most studied endocannabinoids, anandamide, and 2-arachidonoylglycerol (2-AG). Anandamide biosynthetic precursors, the N-arachidonoyl-phosphatidylethanolamines, are produced from the remodeling of phospholipids via the action of N-acyl-transferases (NATs). They are then converted to anandamide, either in 1 step, by N-acyl-phosphatidylethanolamine-selective phospholipase D (NAPE-PLD), or in sequential steps, i.e. by α,β -hydrolase-4 (ABHD4) followed by phosphodiesterase GDE1; or soluble phospholipase A2 (sPLA2) followed by lysophospholipase D (lyso-PLD); or by phospholipase C (PLC) enzymes followed by phosphatases such as PTPN22. The sn-2-arachidonate-containing diacylglycerols serving as biosynthetic precursors for 2-AG are in most cases produced from the action of PLCβ but can also come from phosphatidic acid (PA) via PA phosphohydrolase. However, 2-AG can be also produced from sn-1-lyso-phospholipids via the sequential action of phospholipase A1 (PLA1) and lyso-phospholipase C, or (not shown here) from the dephosphorylation of lysophosphatidic acid. These biosynthetic pathways are shared by anandamide and 2-AG with other N-acyl-ethanolamines and 2-mono-acyl-glycerols, respectively. These congeners, in most cases, do not activate directly the 2 cannabinoid receptors (denoted as CB1 and CB2 here, and as CB1R and CB2R in the main text), but have other targets, some of which shown here, such as orphan G-protein-coupled receptors (GPR55, GPR18, GPR119); the transient receptor potential of vanilloid-type 1 (TRPV1) channel; and peroxisome proliferator-activated nuclear receptors (PPARs). However, also anandamide and, to a lesser extent, 2-AG, have been suggested to be capable of activating some of these targets, particularly TRPV1 and GPR55. Anandamide also inhibits the transient receptor potential melastatin type-8 (TRPM8) channel (blue broken arrow). Both anandamide and 2-AG, following their cellular reuptake by cells, which might be facilitated by a yet-to-be-characterized endocannabinoid membrane transporter (EMT), are inactivated inside cells by enzymatic hydrolysis, respectively by fatty acid amide hydrolase (FAAH) and monoacylglycerol lipase (MAGL). In some cells FAAH, α,β-hydrolase-6 (ABHD6) and, even less frequently, α,β-hydrolase-12 (ABHD12), can also hydrolyze 2-AG. These enzymes are also responsible for the enzymatic hydrolysys of other N-acyl-ethanolamines and 2-mono-acyl-glycerols, respectively, although N-palmitoyl-ethanolamine is preferentially hydrolyzed by N-acyl-ethanolamine acid amidohydrolase (not shown). The 2 endocannabinoids, but not their non-polyunsaturated congeners, can also be oxidized by cyclooxygenase-2 (COX-2), and then processed by prostaglandin synthases, to produce prostamides, in the case of anandamide, and prostaglandin-glycerol esters, in the case of 2-AG. This latter endocannabinoid, via the action of MAGL or ABHD6, can also act as biosynthetic precursor for the nonphospholipase A2-mediated production of prostanoids. Apart from Δ9-tetrahydrocannabinol (THC), plant cannabinoids mentioned in the main text, such as cannabidiol (CBD), CBD acid (CBDA), cannabidivarin (CBDV), cannabigerol (CBG), Δ9-tetrahydrocannabivarin (THCV), THCV acid (THCVA), and CBDV acid (CBDVA), either activate (red solid arrows) or inhibit (red broken arrows) some of the receptors and enzymes of the “endocannabinoidome”. However, they often do so at medium–high micromolar concentrations, and the weight of such interactions in their pharmacology, as compared with others that they have also been suggested to exert, has not yet been fully assessed. Abn-CBD = abnormal cannabidiol; DAG = diacylglycerol; PLCβ = phospholipase β. Adapted from Di Marzo []

Although that depicted above would seem like the natural “evolution” of the definition of the “endocannabinoid system”, things are likely to be even more complicated. First, endocannabinoids, and also cannabinoids, have more molecular targets than just CB1R, CB2R, or thermo-TRPs, and these receptors appear to extend also to proteins that are targeted by other endogenous and exogenous substances. Furthermore, anandamide and 2-AG, like most other lipid mediators, have more than just 1 set of biosynthetic and degrading pathways and enzymes each (Fig. 1), which they often share with “endocannabinoid-like” mediators that may or may not be part of the extended definition of “endocannabinoids” provided above, that is, they may or may not interact with the same proteins to which non-THC cannabinoids bind. In some cases, these degrading pathways and enzymes lead to molecules, such as the prostamides and prostaglandin-glycerol esters (Fig. 1), which are not inactive but instead interact with other receptors, that is, these enzymes are “degrading” for endocannabinoids and “biosynthetic” for other mediators. Finally, some of these enzymes may also have additional completely different functions, for example participate in the chemical modification of molecules that have very little to do with endocannabinoid and cannabinoid targets.

As a result of the above reasoning, some authors now use an extended definition of the endocannabinoid system, such the “enlarged endocannabinoid system”. We, instead, for the sake of clarity, prefer to use the definition of “endocannabinoid-like” mediators [1], that is, mediators belonging to the same chemical class as the endocannabinoids (i.e., amides or esters of long-chain fatty acids), which are not necessarily linked metabolically to anandamide and 2-AG, and have as preferential receptors proteins different from CB1R and CB2R. The ensemble of endocannabinoids, endocannabinoid-like mediators, and their several receptors and metabolic enzymes could then be defined as the “endocannabinoidome” [4]. This is not because its function is necessarily overlapping with that of CB1R and CB2R, but because the genetic or pharmacological manipulation of one member of this new “ome” may ricochet on the activity of these 2 receptors by indirectly influencing the levels or action of anandamide and 2-AG (Fig. 1); hence, the necessity of studying the impact of such manipulations on this system as a whole, by employing the typical methodologies (genomics, transcriptomics, kinomics, metabolomics, lipidomics, etc.) used to investigate the expression and activity of several genes, proteins, and metabolites at once.

Endocannabinoids, Plant cannabinoids and Thermo-TRPs

The first evidence that anandamide and capsaicin, that is, the prototypical plant-derived activator of heat-sensitive TRP channels of vanilloid type-1 (TRPV1), and pungent component of capsicum, have something in common came when some synthetic TRPV1 agonists were found to inhibit the cellular reuptake of anandamide []. Immediately thereafter, anandamide was shown to be the first endogenous agonist of both human and rat TRPV1 channels [], although it was soon clear that this endocannabinoid was not the only lipid mediator capable of exerting this function []. Anandamide was later found also to antagonize TRP channels of melastatin type-8 (TRPM8), which are instead activated by low temperatures and menthol [], whereas the other cold-activated TRP, the TRP channel of ankyrin type-1 (TRPA1), activated by mustard oils, was sensitive only to high micromolar concentrations of the endocannabinoid. The stimulatory, and subsequently desensitizing, effects of anandamide at TRPV1 have been reported in hundreds of studies and represent one of the most important mechanisms, after CB1 activation, through which this lipid mediator exerts its biological functions []. Recently, 2-AG was also shown to activate TRPV1, although at concentrations higher than those required to anandamide to produce the same effect [, ].

Importantly, both THC and non-THC cannabinoids can interact with thermo-TRPs, often in a manner similar to anandamide. The other most abundant, and nonpsychotropic, cannabinoid, cannabidiol (CBD) and, less potently, THCV, cannabigerol (CBG), cannabigerovarin (CBGV), and cannabidivarin (CBDV), activate and desensitize human TRPV1, whereas CBD, CBDV, THCV, and CBG, as well as THC, antagonize rat TRPM8. All these compounds, as well as cannabichromene activate and desensitize rat TRPA1 [, ]. Of these effects, only TRPM8 antagonism is also exerted by cannabinoids in their acid form (which is also the form in which they are naturally produced by the plant) []. Of other proposed heat-sensitive TRP channels, that is TRPV2, TRPV3, and TRPV4, only CBD, CBG, CBGV, THC, and, less potently, THCV, activate and desensitize TRPV2 [], whereas only CBD and THCV are capable of activating and desensitizing rat TRPV3, which is potently desensitized, but not potently activated, by CBGV []. Finally, only THCV is a reasonably good activator of the rat TRPV4, although CBD and CBGV potently desensitize this channel [].

It is important to emphasize that only a few of these effects, which were observed using calcium imaging techniques in intact cells, have been so far confirmed to be due to direct interactions with the channels by using, for example, patch clamp electrophysiology. In particular, this is true to date only for CBD and CBDV activation of TRPV1, TRPV2, and TRPA1 [, ]. Unpublished electrophysiological evidence also exists for the antagonism of TRPM8 (Thomas Voets, personal communication). Nevertheless, the abovementioned thermo-TRPs, especially in the absence of other strong and specific exogenous/xenobiotic modulators, are now considered by all means bona fide “ionotropic cannabinoid receptors”, whereas CB1R and CB2R would thus be defined as “metabotropic cannabinoid receptors” [, , ]. Importantly, several “endocannabinoid-like” mediators, such as, on the one hand, the anandamide congeners N-palmitoylethanolamine, N-oleoylethanolamine, and N-linoleoylethanolamine, as well as several N-acyl-dopamines and N-acyl-taurines, as direct or indirect activators [, –], and, on the other hand, some N-acyl-serotonins, as competitive antagonists [, ], have been shown to interact with TRPV1 in in vitro and in vivo studies.

Other Ways Through Which non-THC Plant Cannabinoids Influence the Endocannabinoid System

Apart from activating (in the case of THC) or antagonizing (in the case of THCV) CB1R and CB2R [], plant cannabinoids have other ways of potentially and indirectly modifying the activity of the endocannabinoid system. Most nonacid plant cannabinoids inhibit, albeit not very potently, the cellular reuptake of anandamide, and CBD is also a moderate inhibitor of anandamide hydrolysis by FAAH [, ], an effect that has been recently reported to also occur in vivo in mice and humans [, ]. If one remembers that several endocannabinoid-like mediators are also inactivated by FAAH (Fig. 1), a consequence of the above findings is that plant cannabinoids can affect the tissue levels of these compounds, too.

None of the cannabinoids tested thus far exerts potent inhibition of 2-AG inactivation by MAGL, although botanical extracts from cannabis varieties producing preferentially either CBG, CBG acid or, particularly, THC acid, as opposed to the pure compounds, do inhibit this enzyme at concentrations < 50 μM, suggesting the presence of MAGL inhibitors among the noncannabinoid components of the extracts []. Conversely, pure CBDV, CBG acid, CBD acid, THC acid, and CBDV acid weakly inhibit (with IC50 values in the 16.6–27.3 μM range) 2-AG biosynthesis by DAGLα []. There are also indirect ways through which CBD (which has extremely low affinity for CB1R and CB2R) inhibits CB1 activity, particularly in the central nervous system, and these have been recently reviewed by McPartland et al. [].

Finally, at high concentrations, THCV behaves as a CB2 agonist, as shown by in vitro and in vivo studies [, , ].

How do Endocannabinoid-like Mediators Influence the Activity of CB1R and CB2R?

Most endocannabinoid-like mediators, that is, those compounds that, according to the definition given above, are not part of the endocannabinoid system but contribute to constitute the endocannabinoidome, do not directly influence the activity of CB1R and CB2R. Whilst some controversial data exist as to the capability of nonarachidonate-containing, polyunsaturated N-acyl-ethanolamines [, ], as well as of the primary amide of oleic acid and sleep-inducing factor, oleamide [], to activate these receptors functionally, some endogenously occurring amides and esters of long chain fatty acids can influence CB1 and CB2 action only indirectly. N-palmitoyl-ethanolamide and, particularly, N-oleoyl- and N-linoleoyl-ethanolamide, by acting as alternative substrates for FAAH [], can retard the degradation of anandamide, and the same is true for some unsaturated N-acyl-amino acids, particularly N-arachidonoyl-glycine [, ], and N-acyl-taurines []. Furthermore, N-palmitoyl-ethanolamide can reduce the expression of FAAH after prolonged exposure of human breast cancer cells []. Likewise, the mono-acyl-glycerols of palmitic, oleic, and linoleic acids enhance the functional activity of 2-AG at CB1R and CB2R, seemingly by inhibiting 2-AG enzymatic hydrolysis [].

Conclusions: The Overlap Between the “Endocannabinoidome” and the “Phytocannabinoidome”

From the evidence reviewed in this article, it is clear that the “endocannabinoidome” and the ensemble of the plant cannabinoids and their molecular targets, which could be defined by analogy as the “phytocannabinoidome”, overlap to some extent. The commonalities between these 2 “omes” are even more striking if one looks at alternative targets (i.e., other than CB receptors and thermo-TRP channels) that have been proposed to date for endocannabinoid-like mediators and phytocannabinoids (see [41] for review). When such targets are receptors, they belong to all 3 major receptor classes, i.e. GPCRs, ligand-sensitive ion channels, and nuclear receptors. In particular: 1) the orphan GPCR, GPR55, has been suggested to act as target for both cannabinoids, that is, THC and CBD, which seem to act as agonist and antagonist, respectively, for this receptor [], and N-palmitoylethanolamine, which seems to be an agonist []. In fact, there is controversial evidence that anandamide and 2-AG may also activate GPR55, and the recent finding of CB1–GPR55 heteromers might explain why some authors have found that the 2 endocannabinoids directly activate this orphan GPCR and most others have not [–]. Another orphan GPCR, GPR18, is instead activated by N-arachidonoyl-glycine and by a synthetic CBD analogue known as abnormal-cannabidiol [, ]. 2) T-type Ca2+ channels have been suggested to be inhibited by both unsaturated long chain fatty acid amides, including some N-acylethanolamines, N-acyl-serotonins, and N-acyldopamines [], and THC and CBD []. 3) peroxisome-proliferator activated receptor-α is activated by both anandamide congeners such as N-palmitoylethanolamine and N-oleoylethanolamine [], which have in this nuclear receptor their preferred target, and some plant and synthetic cannabinoids []; instead, PPARγ is activated, at concentrations of 1–10 μM, by both CBD and anandamide or 2-AG [, ].

However, endocannabinoid-like mediators have additional targets, which they do not share with phytocannabinoids, and the same is true for the latter compounds. Therefore, the overlap between the “endocannabinoidome” and the “phytocannabinoidome” is only partial. Further interactions can be predicted on the basis of the capability of both polyunsaturated endocannabinoid-like mediators and plant cannabinoids to inhibit, or to be oxidized by, cytochrome p450 oxygenases [, ], which, in theory, allows the latter compounds, when administered systemically, to modulate the levels of the former. Finally, CBD acid, and much less so THC acid, were reported to inhibit cycloxygenase-2 [], thus potentially inhibiting the formation not only of prostanoids, but also of prostamides and prostaglandin-glycerol esters. However, such interactions are yet to be demonstrated in vivo. Likewise, most of the interactions between these endogenous and xenobiotic compounds and the targets mentioned in this section have been shown to occur in vitro and their relevance to in vivo pharmacology is yet to be fully clarified. It is, in fact, also the realization that much work still needs to be done to dissect the pharmacological importance of the “endocannabinoidome” and the “phytocannabinoidome”, and, hence, to evaluate fully their and biological/therapeutic relevance, that has convinced many scientists working on this topic to focus so far mostly on the “endocannabinoid system” as it was defined at the turn of the century.

The role of this system and the capability of the most abundant plant cannabinoids, namely THC and CBD, to modulate it in the framework of the treatment of neurological and neuropsychiatric disorders, is the theme of this special issue and of the following chapters.

Electronic supplementary material

Below is the link to the electronic supplementary material.

References

1. Di Marzo V, De Petrocellis L. Fifty years of ‘cannabinoid research’ and the need for a new nomenclature. In: Di Marzo V (ed.) Cannabinoids, Wiley-Blackwell, Oxford, 2014, pp. 261–289.
2. Watkins BA, Kim J. The endocannabinoid system: directing eating behavior and macronutrient metabolism. Front Psychol. 2015;5:1506. doi: 10.3389/fpsyg.2014.01506.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
3. De Petrocellis L, Di Marzo V. Non-CB1, non-CB2 receptors for endocannabinoids, plant cannabinoids, and synthetic cannabimimetics: focus on G-protein-coupled receptors and transient receptor potential channels. J Neuroimmune Pharmacol. 2010;5:103–121. doi: 10.1007/s11481-009-9177-z. [PubMed] [CrossRef] [Google Scholar]
4. Di Marzo V, Wang J (eds). The endocannabinoidome. Academic Press, London, 2015.
5. De Petrocellis L, Bisogno T, Davis JB, Pertwee RG, Di Marzo V. Overlap between the ligand recognition properties of the anandamide transporter and the VR1 vanilloid receptor: inhibitors of anandamide uptake with negligible capsaicin-like activity. FEBS Lett. 2000;483:52–56. doi: 10.1016/S0014-5793(00)02082-2. [PubMed] [CrossRef] [Google Scholar]
6. Zygmunt PM, Petersson J, Andersson DA, et al. Vanilloid receptors on sensory nerves mediate the vasodilator action of anandamide. Nature. 1999;400:452–457. doi: 10.1038/22761. [PubMed] [CrossRef] [Google Scholar]
7. Ahern GP. Activation of TRPV1 by the satiety factor oleoylethanolamide. J Biol Chem. 2003;278:30429–30434. doi: 10.1074/jbc.M305051200. [PubMed] [CrossRef] [Google Scholar]
8. De Petrocellis L, Starowicz K, Moriello AS, Vivese M, Orlando P, Di Marzo V. Regulation of transient receptor potential channels of melastatin type 8 (TRPM8): effect of cAMP, cannabinoid CB(1) receptors and endovanilloids. Exp Cell Res. 2007;313:1911–1920. doi: 10.1016/j.yexcr.2007.01.008. [PubMed] [CrossRef] [Google Scholar]
9. Di Marzo V, De Petrocellis L. Why do cannabinoid receptors have more than one endogenous ligand? Philos Trans R Soc Lond B Biol Sci. 2012;367:3216–3228. doi: 10.1098/rstb.2011.0382.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
10. Zygmunt PM, Ermund A, Movahed P, et al. Monoacylglycerols activate TRPV1—a link between phospholipase C and TRPV1. PLoS One. 2013;8:e81618. doi: 10.1371/journal.pone.0081618.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
11. Petrosino S, Schiano Moriello A, Cerrato S, et al. The anti-inflammatory mediator palmitoylethanolamide enhances the levels of 2-arachidonoyl-glycerol and potentiates its actions at transient receptor potential vanilloid type-1 channels. Br J Pharmacol 2015 Jan 19 [Epub ahead of print]. [PMC free article] [PubMed]
12. De Petrocellis L, Vellani V, Schiano-Moriello A, et al. Plant-derived cannabinoids modulate the activity of transient receptor potential channels of ankyrin type-1 and melastatin type-8. J Pharmacol Exp Ther. 2008;325:1007–1015. doi: 10.1124/jpet.107.134809. [PubMed] [CrossRef] [Google Scholar]
13. De Petrocellis L, Ligresti A, Moriello AS, et al. Effects of cannabinoids and cannabinoid-enriched Cannabis extracts on TRP channels and endocannabinoid metabolic enzymes. Br J Pharmacol. 2011;163:1479–1494. doi: 10.1111/j.1476-5381.2010.01166.x.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
14. De Petrocellis L, Orlando P, Moriello AS, et al. Cannabinoid actions at TRPV channels: effects on TRPV3 and TRPV4 and their potential relevance to gastrointestinal inflammation. Acta Physiol (Oxf) 2012;204:255–266. doi: 10.1111/j.1748-1716.2011.02338.x. [PubMed] [CrossRef] [Google Scholar]
15. Iannotti FA, Hill CL, Leo A, et al. Nonpsychotropic plant cannabinoids, cannabidivarin (CBDV) and cannabidiol (CBD), activate and desensitize transient receptor potential vanilloid 1 (TRPV1) channels in vitro: potential for the treatment of neuronal hyperexcitability. ACS Chem Neurosci. 2014;5:1131–1141. doi: 10.1021/cn5000524. [PubMed] [CrossRef] [Google Scholar]
16. Qin N, Neeper MP, Liu Y, Hutchinson TL, Lubin ML, Flores CM. TRPV2 is activated by cannabidiol and mediates CGRP release in cultured rat dorsal root ganglion neurons. J Neurosci. 2008;28:6231–6238. doi: 10.1523/JNEUROSCI.0504-08.2008. [PubMed] [CrossRef] [Google Scholar]
17. Akopian AN, Ruparel NB, Jeske NA, Patwardhan A, Hargreaves KM. Role of ionotropic cannabinoid receptors in peripheral antinociception and antihyperalgesia. Trends Pharmacol Sci. 2009;30:79–84. doi: 10.1016/j.tips.2008.10.008.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
18. Di Marzo V, De Petrocellis L. Endocannabinoids as regulators of transient receptor potential (TRP) channels: a further opportunity to develop new endocannabinoid-based therapeutic drugs. Curr Med Chem. 2010;17:1430–1449. doi: 10.2174/092986710790980078. [PubMed] [CrossRef] [Google Scholar]
19. De Petrocellis L, Davis JB, Di Marzo V. Palmitoylethanolamide enhances anandamide stimulation of human vanilloid VR1 receptors. FEBS Lett. 2001;506:253–256. doi: 10.1016/S0014-5793(01)02934-9. [PubMed] [CrossRef] [Google Scholar]
20. Ambrosino P, Soldovieri MV, Russo C, Taglialatela M. Activation and desensitization of TRPV1 channels in sensory neurons by the PPARα agonist palmitoylethanolamide. Br J Pharmacol. 2013;168:1430–1444. doi: 10.1111/bph.12029.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
21. Huang SM, Bisogno T, Trevisani M, et al. An endogenous capsaicin-like substance with high potency at recombinant and native vanilloid VR1 receptors. Proc Natl Acad Sci U S A. 2002;99:8400–8405. doi: 10.1073/pnas.122196999.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
22. Movahed P, Jönsson BA, Birnir B, et al. Endogenous unsaturated C18 N-acylethanolamines are vanilloid receptor (TRPV1) agonists. J Biol Chem. 2005;280:38496–38504. doi: 10.1074/jbc.M507429200. [PubMed] [CrossRef] [Google Scholar]
23. Ortar G, Cascio MG, De Petrocellis L, et al. New N-arachidonoylserotonin analogues with potential “dual” mechanism of action against pain. J Med Chem. 2007;50:6554–6569. doi: 10.1021/jm070678q. [PubMed] [CrossRef] [Google Scholar]
24. Verhoeckx KC, Voortman T, Balvers MG, Hendriks HF, Wortelboer MH, Witkamp RF. Presence, formation and putative biological activities of N-acyl serotonins, a novel class of fatty-acid derived mediators, in the intestinal tract. Biochim Biophys Acta. 1811;2011:578–586. [PubMed] [Google Scholar]
25. Thomas A, Stevenson LA, Wease KN, et al. Evidence that the plant cannabinoid Delta9-tetrahydrocannabivarin is a cannabinoid CB1 and CB2 receptor antagonist. Br J Pharmacol. 2005;146:917–926. doi: 10.1038/sj.bjp.0706414.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
26. Bisogno T, Hanus L, De Petrocellis L, et al. Molecular targets for cannabidiol and its synthetic analogues: effect on vanilloid VR1 receptors and on the cellular uptake and enzymatic hydrolysis of anandamide. Br J Pharmacol. 2001;134:845–852. doi: 10.1038/sj.bjp.0704327.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
27. de Filippis D, Iuvone T, d'amico A, et al. Effect of cannabidiol on sepsis-induced motility disturbances in mice: involvement of CB receptors and fatty acid amide hydrolase. Neurogastroenterol Motil. 2008;20:919–927. doi: 10.1111/j.1365-2982.2008.01114.x. [PubMed] [CrossRef] [Google Scholar]
28. Leweke FM, Piomelli D, Pahlisch F, et al. Cannabidiol enhances anandamide signaling and alleviates psychotic symptoms of schizophrenia. Transl Psychiatry. 2012;2:e94. doi: 10.1038/tp.2012.15.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
29. McPartland JM, Duncan M, Di Marzo V, Pertwee RG. Are cannabidiol and Δ(9)-tetrahydrocannabivarin negative modulators of the endocannabinoid system? A systematic review. Br J Pharmacol. 2015;172:737–753. doi: 10.1111/bph.12944.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
30. Pertwee RG. The diverse CB1 and CB2 receptor pharmacology of three plant cannabinoids: delta9-tetrahydrocannabinol, cannabidiol and delta9-tetrahydrocannabivarin. Br J Pharmacol. 2008;153:199–215. doi: 10.1038/sj.bjp.0707442.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
31. Bolognini D, Costa B, Maione S, et al. The plant cannabinoid Delta9-tetrahydrocannabivarin can decrease signs of inflammation and inflammatory pain in mice. Br J Pharmacol. 2010;160:677–687. doi: 10.1111/j.1476-5381.2010.00756.x.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
32. Priller J, Briley EM, Mansouri J, Devane WA, Mackie K, Felder CC. Mead ethanolamide, a novel eicosanoid, is an agonist for the central (CB1) and peripheral (CB2) cannabinoid receptors. Mol Pharmacol. 1995;48:288–292. [PubMed] [Google Scholar]
33. Brown I, Cascio MG, Wahle KW, et al. Cannabinoid receptor-dependent and -independent anti-proliferative effects of omega-3 ethanolamides in androgen receptor-positive and -negative prostate cancer cell lines. Carcinogenesis. 2010;31:1584–1591. doi: 10.1093/carcin/bgq151.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
34. Leggett JD, Aspley S, Beckett SR, D'Antona AM, Kendall DA, Kendall DA. Oleamide is a selective endogenous agonist of rat and human CB1 cannabinoid receptors. Br J Pharmacol. 2004;141:253–262. doi: 10.1038/sj.bjp.0705607.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
35. Bisogno T, Maurelli S, Melck D, De Petrocellis L, Di Marzo V. Biosynthesis, uptake, and degradation of anandamide and palmitoylethanolamide in leukocytes. J Biol Chem. 1997;272:3315–3323. doi: 10.1074/jbc.272.6.3315. [PubMed] [CrossRef] [Google Scholar]
36. Burstein SH, Huang SM, Petros TJ, Rossetti RG, Walker JM, Zurier RB. Regulation of anandamide tissue levels by N-arachidonylglycine. Biochem Pharmacol. 2002;64:1147–1150. doi: 10.1016/S0006-2952(02)01301-1. [PubMed] [CrossRef] [Google Scholar]
37. Bradshaw HB, Rimmerman N, Hu SS, et al. The endocannabinoid anandamide is a precursor for the signaling lipid N-arachidonoyl glycine by two distinct pathways. BMC Biochem. 2009;10:14. doi: 10.1186/1471-2091-10-14.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
38. Saghatelian A, McKinney MK, Bandell M, Patapoutian A, Cravatt BF. A FAAH-regulated class of N-acyl taurines that activates TRP ion channels. Biochemistry. 2006;45:9007–9015. doi: 10.1021/bi0608008. [PubMed] [CrossRef] [Google Scholar]
39. Di Marzo V, Melck D, Orlando P, et al. Palmitoylethanolamide inhibits the expression of fatty acid amide hydrolase and enhances the anti-proliferative effect of anandamide in human breast cancer cells. Biochem J. 2001;358:249–255. doi: 10.1042/bj3580249.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
40. Ben-Shabat S, Fride E, Sheskin T, et al. An entourage effect: inactive endogenous fatty acid glycerol esters enhance 2-arachidonoyl-glycerol cannabinoid activity. Eur J Pharmacol. 1998;353:23–31. doi: 10.1016/S0014-2999(98)00392-6. [PubMed] [CrossRef] [Google Scholar]
41. Alexander S. Common receptors for endocannabinoid-like mediators and plant cannabinoids. In: Di Marzo V, Wang J (eds) The endocannabinoidome. Academic Press, London, 2015, pp. 153–167.
42. Ryberg E, Larsson N, Sjögren S, et al. The orphan receptor GPR55 is a novel cannabinoid receptor. Br J Pharmacol. 2007;152:1092–1101. doi: 10.1038/sj.bjp.0707460.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
43. Martínez-Pinilla E, Reyes-Resina I, Oñatibia-Astibia A, et al. CB1 and GPR55 receptors are co-expressed and form heteromers in rat and monkey striatum. Exp Neurol. 2014;261:44–52. doi: 10.1016/j.expneurol.2014.06.017. [PubMed] [CrossRef] [Google Scholar]
44. Sharir H, Console-Bram L, Mundy C, Popoff SN, Kapur A, Abood ME. The endocannabinoids anandamide and virodhamine modulate the activity of the candidate cannabinoid receptor GPR55. J Neuroimmune Pharmacol. 2012;7:856–865. doi: 10.1007/s11481-012-9351-6.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
45. Kapur A, Zhao P, Sharir H, et al. Atypical responsiveness of the orphan receptor GPR55 to cannabinoid ligands. J Biol Chem. 2009;284:29817–29827. doi: 10.1074/jbc.M109.050187.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
46. Zhao P, Abood ME. GPR55 and GPR35 and their relationship to cannabinoid and lysophospholipid receptors. Life Sci. 2013;92:453–457. doi: 10.1016/j.lfs.2012.06.039. [PubMed] [CrossRef] [Google Scholar]
47. Kohno M, Hasegawa H, Inoue A, et al. Identification of N-arachidonylglycine as the endogenous ligand for orphan G-protein-coupled receptor GPR18. Biochem Biophys Res Commun. 2006;347:827–832. doi: 10.1016/j.bbrc.2006.06.175. [PubMed] [CrossRef] [Google Scholar]
48. Console-Bram L, Brailoiu E, Brailoiu GC, Sharir H, Abood ME. Activation of GPR18 by cannabinoid compounds: a tale of biased agonism. Br J Pharmacol. 2014;171:3908–3917. doi: 10.1111/bph.12746.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
49. Cazade M, Nuss CE, Bidaud I, et al. Cross-modulation and molecular interaction at the Cav3.3 protein between the endogenous lipids and the T-type calcium channel antagonist TTA-A2. Mol Pharmacol. 2014;85:218–225. doi: 10.1124/mol.113.089581. [PubMed] [CrossRef] [Google Scholar]
50. Ross HR, Napier I, Connor M. Inhibition of recombinant human T-type calcium channels by Delta9-tetrahydrocannabinol and cannabidiol. J Biol Chem. 2008;283:16124–16134. doi: 10.1074/jbc.M707104200.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
51. Lo Verme J, Fu J, Astarita G, et al. The nuclear receptor peroxisome proliferator-activated receptor-alpha mediates the anti-inflammatory actions of palmitoylethanolamide. Mol Pharmacol. 2005;67:15–19. doi: 10.1124/mol.104.006353. [PubMed] [CrossRef] [Google Scholar]
52. Sun Y, Alexander SP, Garle MJ, et al. Cannabinoid activation of PPAR alpha; a novel neuroprotective mechanism. Br J Pharmacol. 2007;152:734–743. doi: 10.1038/sj.bjp.0707478.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
53. Scuderi C, Steardo L, Esposito G. Cannabidiol promotes amyloid precursor protein ubiquitination and reduction of beta amyloid expression in SHSY5YAPP+ cells through PPARγ involvement. Phytother Res. 2014;28:1007–1113. doi: 10.1002/ptr.5095. [PubMed] [CrossRef] [Google Scholar]
54. O'Sullivan SE. Cannabinoids go nuclear: evidence for activation of peroxisome proliferator-activated receptors. Br J Pharmacol. 2007;152:576–582. doi: 10.1038/sj.bjp.0707423.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
55. McDougle DR, Kambalyal A, Meling DD, Das A. Endocannabinoids anandamide and 2-arachidonoylglycerol are substrates for human CYP2J2 epoxygenase. J Pharmacol Exp Ther. 2014;351:616–627. doi: 10.1124/jpet.114.216598.[PMC free article] [PubMed] [CrossRef] [Google Scholar]
56. Bornheim LM, Everhart ET, Li J, Correia MA. Characterization of cannabidiol-mediated cytochrome P450 inactivation. Biochem Pharmacol. 1993;45:1323–1331. doi: 10.1016/0006-2952(93)90286-6. [PubMed] [CrossRef] [Google Scholar]
57. Takeda S, Misawa K, Yamamoto I, Watanabe K. Cannabidiolic acid as a selective cyclooxygenase-2 inhibitory component in cannabis. Drug Metab Dispos. 2008;36:1917–1921. doi: 10.1124/dmd.108.020909. [PubMed] [CrossRef] [Google Scholar]
58. Di Marzo V. Targeting the endocannabinoid system: to enhance or reduce. Nat Rev Drug Discov 2008;7:438–455. [PubMed]
Articles from Neurotherapeutics are provided here courtesy of Springer