вход по аккаунту


Quantification of mu and nonЦmu opiate receptors in temporal lobe epilepsy using positron emission tomography.

код для вставкиСкачать
Quantification of Mu and Non-Mu Opiate
Receptors in Temporal Lobe Epilepsy Using
Positron Emission Tomography
Helen S. Mayberg, MD,"? Bernard Sadzot, MD," Carolyn Cidis Meltzer, MD," Robert S. Fisher, MD, PhD,?
Ronald P. Lesser, MD,? Robert F. Dannals, PhD,"S John R. Lever, PhD,"S Alan A. Wilson, PhD,"
Hayden T. Ravert, PhD," Henry N. Wagner, Jr, MD,"S R. Nick Bryan, MD, PhD,"
Christina C. Cromwell, BA,$ and J. James Frost, MD, PhD"$§
Alterations in a variety of neurotransmitter systems have been identified in experimental models of epilepsy and in
brain tissue from patients with intraaable temporal lobe seizures. The availability of new high-affinity radioligands
permits the study of some neuroreceptors in vivo with positron emission tomography (PET).We previously characterized the in vivo binding of "C-carfentanil, a potent and selective mu opiate receptor agonist, and described increases
in "C-carfentanil binding in the temporal neocortex of patients with unilateral temporal lobe epilepsy. These studies
have been extended to "C-diprenorphine, which labels mu, kappa, and delta opiate receptor subtypes. Paired measurements of opiate receptor binding were performed with PET using "C-carfentanil and "C-diprenorphine in patients
with unilateral temporal lobe seizures. Carfentanil binding, reflecting changes in mu opiate receptors, was increased
in the temporal neocortex and decreased in the amygdala on the side of the epileptic focus. Diprenorphine binding,
reflecting mu as well as non-mu opiate subtypes, was not significantly different among regions in the focus and
was decreased in
nonfocus temporal lobes. Regional glucose metabolism, measured using '*F-2-fluoro-2-deoxyg1ucose,
the mesial and lateral aspects of the temporal lobe ipsilateral to the epileptogenic focus. The variation in pattern of
carfentanil and diprenorphine binding supports a differential regulation of opiate subtypes in unilateral temporal
lobe epilepsy.
Mayberg HS, Sadzot B, Meltzer CC, Fisher RS, Lesser RP, Dannals RF, Lever JR, Wilson AA, Ravert HT,
Wagner HN Jr, Bryan RN, Cromwell CC, Frost JJ. Quantification of mu and non-mu opiate receptors in
temporal lobe epilepsy using positron emission tomography. Ann Neurol 1991;30:3- 11
A variety of neurotransmitters and their associated receptors have been implicated in the initiation and termination of seizures in humans 11-91. A specific role
for opiate receptors and endogenous opioid peptides
in seizure mechanisms has become increasingly well
recognized. The cumulative evidence suggests that central nervous system (CNS) opioid peptides are involved in the termination of seizures and may, in fact,
serve a primary role as an endogenous anticonvulsant.
This conclusion is supported by experimental models
of epilepsy in several animal species 110-231. Induced
generalized seizures are associated with the release of
an anticonvulsant substance into the cerebrospinal fluid
(CSF) that contains both a beta-endorphin and a metenkephahn precursor and whose anticonvulsant effects
are reversed by naloxone [IS, 161. Interestingly, a recent study also showed increased CSF levels of en-
kephalin in epilepsy patients [17}. Other studies suggest that the anticonvulsant effects of both peptide and
nonpeptide opioids are mediated through mu, delta,
and kappa opiate receptors [18-2 13. Specifically, anticonvulsive effects of pala-D-leu-enkephahn and dynorphin, acting respectively at delta and kappa receptors,
have been demonstrated C20). In addition, anticonvulsant synthetic opioids with selective mu, delta, and
kappa properties have been synthesized 12 I}. Consistent with these findings, chronically induced seizures
in rodents have been shown to specifically alter mu
opiate receptors C22, 231. Changes in delta or kappa
receptor density or affinity, however, have not yet
been demonstrated.
Over the last several years, with the availability of
new high-affinity radioligands, we have developed and
validated in vivo imaging methods to measure both mu
From the *Department of Radiology, Divisions of Nuclear Medicine
and Neuroradiology, and the Departments of ?Neurology, SEnvironmental Health Sciences, and §Neuroscience, The Johns
University School of Medicine, Baltimore, MD.
Received Oct 5 , 1990, and in revised form Dec 19. Accemed for
publication Dec 22, 1990.
Address correspondence to Dr Frost, Johns Hopkins University
School of Medicine, DeDartment of Radiolom. Room B1-130. 600
North Wolfe Street, Baltimore, MD 21205.-'
Copyright 0 1991 by the American Neurological Association
and non-mu opiate receptor binding in patients and
normal volunteers using positron emission tomography
(PET) r24-291. T w o opiate receptor ligands have been
used in these studies: the mu-selective opiate agonist
"C-carfentanil (CFN), and the opiate partial agonist
"C-diprenorphine (DPN), which has similar affinity
for mu, delta, and kappa opiate receptors. These two
ligands were previously evaluated independently in
normal volunteers to quantify their hnetic binding
properties and determine regional receptor densities
and affinities 125-27). In addition, paired studies of
CFN and DPN performed in the same normal volunteers {28) showed differential regional binding comparable to the known distribution of m u and non-mu
opiate receptor subtypes mapped in vitro r30, 31).
CFN preferentially binds to sites in the amygdala, cingulate cortex, thalamus, and layers I11 and IV of the
cerebral cortex {24, 28, B. Sadtot, unpublished data,
1988}, all of which are areas with high concentrations
of m u opiate receptors {3U, 31). DPN binding, o n the
other hand, has a more homogeneous distribution that
is consistent with autoradiographic maps demonstrating delta receptors in the basal ganglia and kappa receptors in the amygdala and deep layers of the cortex,
particularly layers V and VI 128, 30-331. CFN PET
studies in patients with complex partial epilepsy have
identified increases in m u opiate receptor binding in
the temporal neocortex ipsilateral to the electroencephalographic (EEG) focus {29J This increase in opiate receptors was inversely related to the decrease in
regional glucose metabolism present in specific regions
of involved temporal lobe ncocortex.
T h e present study was initiated to examine potential
differences in m u and non-mu opiate receptor binding in temporal lobe epilepsy. Interictal paired PET
measurements were made using CFN and DPN in patients with unilateral complex partial seizures. Regional
cerebral glucose metabolism (CMRGlc) using I8F-2fluoro-2-deoxyglucose (FDG) was also measured in
each patient. We hypothesized that DPN binding, reflecting changes in both m u and non-mu (delta and
kappa) opiate receptors, would differ from CFN binding (reflecting a change in m u receptors alone) if m u
and non-mu receptors were affected differently in epilepsy.
Identification of Epileptogenic Foci
Standardized criteria were used to establish the hemisphere
of origin of the epileptogenic focus in all patients. Patients
were admitted to the Epilepsy Monitoring Unit in the Department of Neurology at the Johns Hopkins Hospital for
diagnostic evaluation. A standardized EEG montage of 30 to
34 electrodes arranged in a modified International System
array, designed for identification of the hemisphere and lobe
of onset of partial complex seizures, was used to evaluate all
patients 1351. Video monitoring and clinical observation were
performed continuously over a 4- to 7-day period. Hemisphere of onset of each patient's seizure was established using
the following criteria: (1) consistent interictal spiking (when
present) over either temporal area in the scalp recording and
(2) local spiking or focal rhythmic activity at the start of a
clinical seizure. Unilateral temporal lobe epilepsy was diagnosed if a minimum of three ictal EEG tracings demonstrated
seizure onset from one temporal lobe and there was no evidence of a contralateral focus. Interictal lateralization was
established in 7 of the 11 patients. Unambiguous unilateral
ictal onset was identified in 10 of the 11 patients. The remaining patient had 10 seizures recorded, with 7 of the 10
identifying a consistent unilateral focus.
Selection of the PET Imaging Plane
Materials and Methods
Patient Characteristics
Subjects for this study were screened by two of us (R.S.F.,
and R.P.L.), from a cohort of epilepsy patients who are regularly treated in the Neurology Clinic at the Johns Hopkins
Hospital. Only patients with complex partial seizures and a
unilateral temporal lobe focus, identified from ictal scalp EEG
obtained during continuous inpatient video monitoring, were
eligible. All patients had poorly controlled seizures despite
appropriate plasma levels of anticonvulsant medications. All
patients were being evaluated for surgical treatment of their
4 Annals of Neurology
epilepsy. The inclusion criteria for participation in the PET
imaging study were (1) complex partial seizures, as defined
by the International Classification of the Epilepsies 1341; (2)
a unilateral temporal seizure focus documented by a minimum of three consistent ictal EEG tracings recorded during
an evaluation in the epilepsy monitoring unit; and (3) no
underlying structural lesion evident by x-ray computed tomography (CT).
Eleven patients, 7 men and 4 women, ages 20 to 40 years,
met all of these inclusion criteria. The mean age at onset of
seizures in the 11 patients was 10 2 5.5 years, with an aver9.0 years. Seizure frequency
age illness duration of 21
ranged from 1 to 30 seizures per month. All patients were
classified as having idiopathic epilepsy. Three patients had a
history of febrile seizures prior to age 2, and 1 patient had
documented perinatal hypoxia. The average intelligence quotient in this group of patients was 93 & 14, and total years
of education was 14 t 2 years. All patients were caking one
or more of the following anticonvulsant medications at the
time of the PET studies: carbamazepine, phenytoin, primidone, valproic acid, or clotazepate. Informed consent, using
guidelines established by the Johns Hopkins University
School of Medicine Joint Committee on Clinical Investigation, was obtained from each subject prior to all imaging
Vol 30 N o 1 July 1991
Prior to the acquisition of the three PET studies, CT was
performed with or without magnetic resonance imaging
(MRI) in all patients to select the imaging plane to be used
for the PET scans. This procedure also ensured symmetrical
alignment and comparable positioning among all subjects.
The anatomical imaging study was used to identify a set of
three imaging planes: 32 mm apart (the separation of the
three planes of the Neuro ECAT PET scanner [CTI, Knoxville, TN)), parallel to the long axis of the temporal lobe,
and passing through the amygdala and occipital cortex in the
Fig I . PET imaging plane and region of interest selection. Imaging planes for the three PET studies in each subject were selected using x-ray CT with or withoat MRI. (Aj A plane offset
20 degrees from the glabellar-inion line, passing through the
amygdala, and trauersing the long axis of the temporal lobe was
identifed using x-ray CT. In a subset of patients, a comparable
plune was first identij5ed on an MRI sagittal plane where the
amygdala was clearb visible. (Bi A plane, bisecting the amyg&la and passing through the long axis of the temporal lobe and
occipital lobe, was selected. (C) The oblique transaxiaI slice at
this leuel is seen. A CT study was then pevfmed to confirm the
positioning. (0)Regions of interest were identiJied directly from
the ”C-cavfentanilPET scan rising the corresponding axial CT
or MRI as a guide. A = amygdda; 1, 2 , and 3 = anterior
temporal cortex; 4, 5 , and 6 = mid temporal cortex; 7, 8, and
9 = posterior temporal cortex; 0 = occipital cortex.
most caudal plane. A thermoplastic face mask (Polysplint,
Poly-Med Manufacturing, Baltimore, MD) with an alignment
line drawn o n the surface of the mask at the time of the CT
or MRI scan was used to ensure reproducible positioning
between the anatomical and functional imaging studies.
Using CT alone (5 patients), the imaging angle in each
patient was defined by a plane -20 degrees from the
glabellar-inion line (parallel to the bicommissural line [36])
identified on a lateral x-ray topogram. Serial 4-mm-thick
scans were then acquired to identify a plane passing through
the center of the amygdala and long axis of the temporal lobe
and bisecting the occipital cortex [37], an area with low levels
of opiate receptors, which is used in the quantification of
both CFN and DPN binding (Fig IA).
In 6 patients, the imaging plane was selected using x-ray
CT in conjunction with MRI and a recently developed
method [38]. In brief, an external localizing device, consisting of a series of paramagnetic tubes visible on MRI, was
affixed to an individually fitted thermoplastic mask. A plane
of interest, passing through the center of the amygdala and
the long axis of the temporal lobe, bisecting the occipital
lobe, was identified on a parasagittal MRI at approximately
15 mm lateral to the midline (Fig IB). The plane was defined
by its angle and point of intersection on the localizing device
and its relationship to the “landmark” or reference position
of the MRI scanner. The position of the plane (Fig 1C) was
confirmed using x-ray CT and then marked on the mask
using a calibrated alignment laser. This method was shown
in phantom studies to be accurate and reproducible to within
1 mm and 1 degree of the desired plane [38).
MRI studies were also performed in all patients as part of
the clinical evaluation for possible epilepsy surgery. These
clinical scans were acquired using a standardized protocol. A
sagittal T1-weighted series (5-mm slice thickness, 1.5-mm
gap; TR 600ITE 20), axial proton density, and T2-weighted
series (5-mm thickness, 2.5-mm gap; TR 3OO/TE 30-100)
and a T1 coronal series (TR 834iTE 30) were obtained with
and without gadolinium contrast enhancement. A neuroradiologist (R.N.B.), blind to both the clinical status of the patients and the results of the PET studies, reviewed the l l
patient scans and scans from 11 age-matched normal control
Mayberg et al: Opiate Receptors in Epilepsy 5
subjects in random order. MRI scans were rated as normal
or abnormal, and the side and location of any identified abnormality were noted.
Carfenlanil, Dz$renorphine, and Fluomdeoxyglucose
PET Studies
Two opiate PET studies (DPN followed by CFN) were performed in each patient on a single day, with a 60-minute rest
period between scans. CFN (20 mCi dose; average mass,
0.05 mgikg; specific activity, > 2,500 Ci/mmol; maximal receptor occupancy, < 5%) and D P N (20 mCi dose; average
mass, 0.05 mg/kg; specific activity, > 2,500 Ciimmol; maximal receptor occupancy, < 5%) were synthesized using published methods 133, 40). Brain radioactivity of each tracer
was measured over 90 minutes using scan times of increasing
duration (2 to 2 1 minutes). Scans were acquired in the highresolution mode (full-width half-maximum [FWHM), 0.8
cm), corrected for radioactive decay, and attenuation using a
manually placed ellipse around each brain slice and an attenuation coefficient of 0.088. Scans were then summed over the
interval of 35 to 70 minutes for CFN, and 50 to 90 minutes
for DPN, and smoothed using an 8 x 8 weighted filter to a
final resolution of 13 mm. Scans are routinely smoothed using this method to reduce noise. The patient was positioned
for each study using the MRI or CT localization line drawn
on the thermoplastic mask and a laser light coincident with
the PET scanner’s detector rings. The patients’ head position
within the scanner was monitored and continually adjusted
throughout each PET study by a PET technologist to ensure
maintenance of the preselected plane of interest.
Interictal CMRGlc was also measured in each of the 11
patients using standard methods [29, 41). Interictal FDG
scans were performed within 7 days of the paired CFN and
DPN studies (usually on the following day). Scans were acquired parallel to and including the three preselected planes
used in the paired opiate studies, on which regions of interest
were sampled. Repositioning was achieved using the alignment line previously drawn on the thermoplastic mask. A
total of 12 slices covering the full rostral-caudal extent of the
brain were obtained. Regional CMRGlc was calculated using
published rate constants [ 4 17, and normalized to the average
whole-brain gray matter in each subject. Normalized regional
values (rCMRGlc) were used in all statistical anlayses.
Paired regions of interest (4 x 4 pixels) were drawn symmetrically on the CFN PET image in the amygdala, temporal
neocortex, and occipital cortex of each hemisphere using the
coincident CT or MRI as a guide (see Fig 1D). Regions of
interest were first placed on the CFN study (see Fig 1D).
Identical regions were then placed on corresponding DPN
and EDG images. Temporal neocortex regions were of equal
size and covered the entire anterior-posterior extent of the
cortex present in the lowest PET slice (see Fig 1D regions 1
to 9).
EEG was not monitored during the scanning period, but all
patients were continuously observed for evidence of typical
seizures. Patients and their accompanying relatives were interviewed prior to and following each PET scan to determine
the occurrence of temporally relevant epileptic phenomena.
The average time between the patient’s most recent seizure
and the PET studies was 8 2 8 days. Only 1 patient had a
seizure on the same day as the PET studies.
6 Annals of Neurology Vol 30 No 1 Juiy 1991
Receptor Binding Quantification
Specific to nonspecific binding ratios (sins) were computed
for each study in all regions using the (region - occipital)/
occipital ratio. These s/ns values were used for subsequent
statistical analysis. A linear relationship between the regionoccipital ratio 35 to 70 minutes after injection and the binding potential (k3/k4) has been demonstrated previously using
compartmental modeling analysis of paired high-specificactivity CFN studies (without and with naloxone) in normal
volunteers 1251. A linear relationship between total binding
(50 to 90 minutes after injection) and k3ik4 or receptor
has also been calculated from paired high- and
density (B,)
low-specific-activity D P N studies (26, 27, J. Price, unpublished data, 1990). Simulation studies established that these
receptor measurements are related to receptor number or
affinity and not to other properties such as blood flow or
blood-brain permeability [25, 26, 291. Nonspecific binding
was estimated directly from the occipital cortex in the CFN
scans, since in vivo competition studies with naloxone [25]
and in vitro autoradiography [30, 31) confirmed the virtual
absence of mu opiate receptors in this brain region. With
DPN, nonspecific binding was represented by 30% of the
total DPN activity in the occipital cortex. This value was
determined from competitive blocking studies performed in
a set of normal volunteers who were pretreated with a known
opiate receptor saturating dose of naloxone (which binds with
similar affinity to mu, delta, and kappa subtypes) followed by
the administration of high-specific-activity DPN 1281. Seventy percent of the occipital binding was displaced by naloxone and, therefore, was assumed to be bound to non-mu
opiate receptors since there are negligible numbers of mu
receptors in this region.
Data Analysis
Statistical analyses were designed to assess if focal as well as
lateralized differences in opiate binding and glucose metabolism were present in the known epileptogenic temporal lobe.
Regional analyses were confined to data obtained from the
imaging plane passing through the amygdala, temporal lobe,
and occipital lobe. Twelve pairs of identical regions were
identified in the focus and nonfocus hemispheres of each
patient for each of the three PET studies (CFN, DPN, and
FDG). Cortical regions of interest in each hemisphere (see
Fig 1D) were grouped into anterior (regions 1 to 3), middle
(regions 4 to G), and posterior (regions 7 to 9) temporal
regions. A single region within each amygdala was also sampled. Nonspecific binding was estimated using the occipital
cortex. A specific-to-nonspecific binding ratio (sins) for each
region in the epileptogenic (focus) and contralateral (nonfocus) hemisphere was calculated for both CFN and DPN.
Focus and nonfocus binding or metabolism (rCMRGlc) were
compared for each tracer independently, using repeated measures methodology. Analysis of variance (ANOVA) was performed using two intrasubject variables: hemisphere (focus
side and nonfocus side) and region (anterior temporal, mid
temporal, posterior temporal, and amygdala). Intrasubject
differences in the degree of binding asymmetry were assessed
by comparing the asymmetry ratios for the three studies
(CFN and D P N specific binding focus-nonfocus ratio or
rCMRGlc focus-nonfocus) using analysis of variance with two
repeated variables per subject: tracer (CFN, DPN, and FDG)
PET Quant&wtzon of Regional Opiate Receptor Binding and Glucose Metabolism
in Patients with Unilateral Temporal Lobe Epilepsy
slns Binding
slns Binding
Anterior temporal
cortex (Tl-3)
cortex (T4-6)
Posterior temporal
cortex (T7-9)
* 0.53'
* 0.45
2.03 2 0.57
1.61 f 0.52
5.38 2 1.48
5.52 & 1.77
5.81 t 1.86
5.45 ? 1.77
0.73 2 0.19
0.77 -t 0.18
* 0.18
* 0.17
5.96 f 1.87
5.74 2 1.72
0.91 t 0.17
1.60 t 0.55d
1.30 i 0.61d
1.08 f 0.59
* 1.70
* 1.39
1.11 t 0.24
"Focus > nonfocus, all regions p = 0.035.
bFocus < nonfocus, all regions p = 0.02.
'Focus < nonfocus, p < 0.05.
dFocus > nonfocus, p < 0.01.
specific-nonspecific binding ratio; rCMRGlc
glucose metabolic rate; region-whole brain.
and region (anterior temporal, mid temporal, posterior temporal, and amygdala). Tukey's post hoc analyses were used
to localize those regions showing significant differences as
determined by ANOVA. The relationships between opiate
binding and clinical, MRI, and demographic variables were
assessed using Pearson product-moment correlation coefficients.
There were significant differences among the results
from the CFN, DPN, and FDG studies. The greatest effect was seen in the temporal neocortex where
CFN binding was significantly higher on the side of
the EEG focus when compared with the nonfocus side
(ANOVA F(1,lO) = 5.95, p = 0.035). Furthermore,
significant regional differences between focus and nonfocus CFN binding were found (side-by-region interaction: F(3,30) = 13.00, p = 0.0001) (Table). Tukey's
post hoc tests demonstrated significant increased binding in the mid and posterior temporal regions on the
side of the EEG focus (p < O.Ol), and significantly
decreased binding in the amygdala on the same side (p
< 0.05). There was no significant difference between
focus and nonfocus binding in the anterior part of the
temporal cortex.
DPN binding, on the other hand, showed no significant difference between the focus and nonfocus
hemispheres (ANOVA: F(1,lO) = 0.037, p = 0.85).
Although a significant side-by-region interaction was
found (F(3,SO) = 3.88, p = 0.02), post hoc tests failed
to detect significant differences between paired regions
in the focus and nonfocus temporal lobes (see Table).
Regional glucose metabolism was significantly decreased on the side of the EEG focus compared to the
nonfocus side (ANOVA F(1,lO) = 8.41, p = 0.02),
but unlike CFN, there were no significant differences
Fig 2. "C-carfentanil (CFN), "C-diprenorphine (DPNj, and
'XF-2-JElkoro-2-deoxyglucose(FDG) PET, imuies in u 26-yearold. right-handed woman with left-sided temporal lobe seizures.
In the CFN study (left),a 70% increuse in mu opiate receptor
binding in the mid and posterior temporal neocortex und a 40%
decrease in binding in the amygdala are present in the lej5t temporal lobe ipsilateral to the EEG focus.In the DPN study (center), there is a 15% increase in binding in the lt$t temporal
neocortex compared t o the contrulateral side, but this increase
was not statistically signzficant in the analysis of all patients.
In the FDG study (right), there is dyfuse hypometabolism, approximateiy 3076, involving the unzygdala and temporal neocortex in the electrically abnormal IeJt temporal lobe.
among individual regions (side-by-region interaction:
F(3,30) = 1.86, p = 0.16). In other words, diffuse
hypometabolism was present throughout the temporal
lobe on the side of the EEG focus (see Table).
These quantitative changes in mu and non-mu opiate receptor binding and regional metabolism on the
side of the seizure focus can be visualized in the set of
PET images taken in a single patient (Fig 2).
Mayberg et al: Opiate Receptors in Epilepsy 7
binding was not correlated with either the magnitude
of change in D P N binding or the degree of glucose
hypometabolism. In other words, patients with the
greatest increase in CFN binding did not necessarily
demonstrate the greatest change with DPN or FDG.
There were no differences in any of the three studies
between patients with early onset of seizures, defined
as onset at age 10 or younger, and patients with onset
after age 10. Similarly, there were no significant correlations between binding and patient age, seizure frequency, onset age, duration'in years, or time from the
last seizure.
C-11 diprenorphine
n cn
ant. temp.
mid. temp.
post. temp.
Fig 3.Regional asymmetv of glucose metabolism and opiate receptor binding. Regional asymmetry (mean SO) was measured
in all patients In = 11) using the ratio of receptor binding or
metabolism in the electricalb identified abnormal hemisphere
(focus) to the contralateral hemisphwe (nonfocus)for each of the
three tracers. Signafcant hypometabolism was present in all regions i n the '8P-2~uom-2-deoxyglucose(FDG)studies (focusnonfocus ratio < I ) . There was no signi3cant asymmetry in
"C-diprenorphine (DPN) binding in any of the measured regions. Significant increases in mu receptors on the side of the focus were present in the "C-carfentanil (CFNI studies (focusnonfocus ratio > I j and the magnitude o f regional a.fymmetv
was significanti) greater than that measured with both DPN
(*mid temporal, p < 0.01, posterior temporal, p < 0.01) or
FDG ($posteriortemporal, p < 0.OSJ.
Intrasubject comparisons of CFN, DPN, and FDG
asymmetry also identified significant differences
(F(2,20) = 5.66, p = 0.011). Specifically, asymmetry
(focus-nonfocus ratio) was significantly greater for
CFN than D P N across all regions (mean focus-nonfocus ratio for CFN = 1.15 k 0.32, D P N = 1.02
-+ 0.14; F(1,lO) = 11.56, p = 0.007), and was most
prominent in the mid and posterior temporal neocortex (study-by-region interaction: F(3,30) = 11.2, fi =
0.0001; Tukey's post hoc p < 0.01) (Fig 3). When
CFN and FDG asymmetries were compared, a significant difference among regions was demonstrated
(study-by-region interaction: F(3,30) = 3.95, p =
0.018); greater asymmetry was present in the posterior
temporal cortex measured with CFN than with FDG
(Tukey's post hoc p < 0.05) (see Fig 3). Comparisons
of CFN and FDG asymmetry across all regions demonstrated that overall asymmetry tended to be greater for
CFN (focus-nonfocus ratio = 0.24 f 0.25) than for
FDG (focus-nonfocus ratio = 0.15 f 0.11) but this
difference was not statistically significant (F(1 , l O ) =
3.42,p = 0.09).
Despite clear intrasubject differences among the
three tracers, the magnitude of the increase in CFN
8 Annals of Neurology Vol 30 KO 1 July 1991
Clinical MRI was performed in all 11 subjects. Seven
of the 11 studies were normal. The remaining 4 had
identifiable abnormalities not visible on x-ray CT.
Three of the 4 patients had an MRI-detected lesion
on the side of the known epileptogenic focus in the
temporal lobe. One patient had a 5-mm porencephalic
cyst identified on T1- and T2-weighted, and proton
density scans. The second patient had a left medial
temporal lesion seen on T2 and proton density scans,
compatibile with gliosis. The third patient had a 10-mm
temporal cavernous angioma with evidence of an old
hemorrhage. The fourth patient had an incidental abnormality in extratemporal white matter in the ipsilateral hemisphere.
Patients with MRI abnormalities did not show significant differences in binding or metabolism compared
to patients with normal MRI scans.
In the present study, paired PET measurements of opiare receptors were performed using CFN and D P N in
patients with unilateral temporal lobe seizures. CFN
binding was increased in temporal neocortex and decreased in the amygdala on the side of the epileptic
focus, reflecting changes in mu opiate receptors. DPN
binding, which measures mu and non-mu opiate receptor binding, was not significantly different among
regions in the focus and nonfocus temporal lobes. Regional glucose metabolism was decreased in the mesial
and lateral aspects of the temporal lobe, ipsilateral to
the epileptogenic focus, similar to previous reports
We confirm in this study our previous observation of
increased mu opiate receptor binding in the temporal
cortex ipsilateral to the known epileptogenic focus using CFN. As before, the area of greatest increase in
opiate receptor binding was in the cortex, rather than
in the amygdala and hippocampus, which have been
presumed to be primary sites of epileptogenic foci
1441. In light of the known primary role of the opioid
peptides and receptors in seizure termination, we suggest that mu receptor increases in neocortex may be a
diffuse compensatory response to locally propagated
seizures and interictal spike activity. Conversely, the
associated decrease in mu receptor binding in the
amygdala may reflect nonspecific tissue damage or
downregulation in the region of the presumed epileptogenic focus. There were no significant differences
between the two hemispheres in amygdala volumes
determined from the clinical MRI scans, arguing that
the decreased amygdala binding may in fact be due to
In an effort to identify clinical parameters that might
relate to these receptor changes, receptor binding was
correlated with measures related to illness duration and
severity. These correlations were unrevealing. Since
the endogenous opiate system appears to play a predominant role in seizure termination, it will be important in future studies to evaluate the relation between
the duration of individual seizures and increases in mu
opiate receptor binding. Patients with a history of status epilepticus would be of particular interest and
might actually show reduced mu opiate receptor binding. Furthermore, studies using a multislice scanner
and precise registration of PET studies with cortical
grid electrode maps of ictal, postictal, and interictal
spike activity will allow for direct examination of opiate
receptor changes in relation to focal abnormalities in
electrical behavior. Secondly, all patients were studied
while taking anticonvulsant medications. The chronic
effects of anticonvulsants on regional opiate receptor
number, affinity, or occupancy have not been evaluated. While it is unlikely that specific drugs would have
lateralized effects on regional opiate binding, medications might affect the overall magnitude of binding.
This will need to be addressed in separate experiments.
An unexpected finding from this study was the absence of a significant change in DPN binding in the
presence of an increase in mu receptor binding in the
same patient. This may be explained by one of several
mechanisms. One possibility is that either delta or
kappa receptor number or affinity is decreased in the
neocortex of the temporal lobe, counteracting the mu
receptor increase. This would most likely be a kappa
receptor effect since delta receptors are normally low
in number in the cortex.
Alternatively, endogenous delta and/or kappa opioid
peptides may be released, occupying delta or kappa
receptors, and hence lowering the number of available
binding sites, similar to findings in animals under stress
C45). Either of these mechanisms could produce an
attenuated D P N signal. Unfortunately, the amount of
endogenous opioid peptide bound to the opiate receptors cannot be directly measured by in vivo PET methods and, therefore, changes in receptor number cannot
be distinguished from changes in occupancy. No characteristic seizures, however, were observed during any
of the PET studies and the single patient with a seizure
within the hour prior to the PET study actually had
greater binding than the group mean. It is unlikely,
therefore, that this latter mechanism would account for
our findings. Accordingly, we believe a reduction in
delta or kappa receptor affinity or number is the most
likely explanation for our results.
Methodological issues related to the use of ligands
with affinity for multiple receptor subtypes, as was employed in this study, require comment since individual
receptor subtypes (i.e., mu, delta, or kappa) cannot be
distinguished. In addition, despite relatively discrete
localization of mu, delta, and kappa opiate receptors in
different brain regions, most regions contain at least a
small fraction of each receptor subtype, further complicating interpretation of changes. However, the known
high concentration of kappa and mu receptors in temporal cortex (Compared to delta receptors) would suggest that changes in these two subtypes account for the
findings of this study. This evidence is indirect, since
in spite of clear pharmacological evidence in animal
models of the anticonvulsant effects of delta and kappa
agonists, selective changes in delta or kappa opiate receptor number or affinity have not been evaluated. Decreases in prodynorphin mRNA and increases in proenkephalin mRNA in entorhinal cortex-hippocampal
regions were recently reported following both electroconvulsive shock and perforant pathway stimulation
in the rat r46, 47). These data further demonstrate a
differential regulation of the various opioid systems
in seizure mechanisms. Therefore, when nonparallel
changes among receptor subtypes occur (mu increases
with concurrent kappa or delta decreases), detection of
differences between the focus and nonfocus hemisphere will be limited when a nonselective ligand like
DPN is used, and will falsely lead to the conclusion
that no changes in opiate receptors exist.
This hypothesis is supported by examination of individual D P N scans where D P N binding was actually
decreased on the focus side in 3 of the 11 patients.
These same patients, however, had consistent increases
in CFN and decreases in FDG. Thus, these two tracers
correctly identified the epileptogenic focus, despite the
contradictory DPN findings. The occurrence of opposing changes in kappa and mu receptors could also account for the absence of detectable asymmetry in opiate binding in seizure patients recently reported using
"F-cyclofoxy [48}, a nonselective opiate antagonist
with similar affinity for mu and kappa receptors, and
low affinity for the delta subtype 1431. Findings from
the cyclofoxy study, as well as the CFN and D P N studies presented here, are not inconsistent with a change
in both mu and nun-mu receptors on the side of the
focus, but the current methods preclude direct quantification of changes in the non-mu subtypes. This inability to separate receptor subtypes is not unique to
studies of epilepsy, and emphasizes the need for devel-
Mayberg et
Opiate Receptors in Epilepsy 9
opment and use of subtype selective tracers. Future
imaging studies using a kappa selective ligand may help
to clarify the hypothesized decreases in this subtype
suggested by the DPN studies.
PET studies performed in patients prior to epilepsy
surgery have several advantages over postsurgical in
vitro analyses of temporal lobe tissue [l, 5-9, 501.
Although only a single receptor can be evaluated at
any one time with PET, and sequential studies are required to evaluate more than one receptor type, comparisons of brain regions within, adjacent to, and remote from the known electrical focus can be made
simultaneously. Similarly, the area of seizure origin can
be directly compared with the homotopic region in the
electrically less-involved contralateral hemisphere and
with data from normal subjects. One can address laterahation and localization within a lobe or hemisphere.
In addition, one can assess the relationship between
regional receptor changes and localized electrical spike
activity measured in chronically monitored patients, as
well as observe in vivo pharmocokinetics, not possible
with other approaches. Analysis of presurgical PET
data is not confounded by assumptions of what is “normal” and “abnormal” brain, or by anesthesia, effects of
tissue fixation, and structural artifacts introduced by
the surgical resection itself-variables unique to a resected tissue block.
In vitro studies of multiple receptors in resected human tissue and preclinical animal models remain the
gold standard in the identification of transmitter and
peptide systems involved in seizures, and are important
tools in the selection development of PET ligands for
human studies [ S O } . This strategy has been applied to
few receptors other than the opiate receptor, as PET
tracers for many of the receptors implicated in epilepsy
await development. Decreases in central benzodiazepine receptors, using the tracer “C-flumazenil, have
been demonstrated in temporal and extratemporal foci
anatomically coincident with epileptic spike activity
I5 11. Future studies of excitatory neurotransmitter receptors are anticipated and are particularly relevant to
seizure initiation.
In conclusion, this study demonstrates that changes
in opiate receptor binding due to changes in receptor
affinity, number, or occupancy occur in the temporal
neocortex adjacent to the epileptic focus. The findings
also demonstrate that the neuroreceptor changes are
subtype-specific. Subtype-specific neuroreceptor imaging by PET permits elucidation of neurochemical
features of human epilepsy and, in the future, may aid
in the individualized planning of surgical treatment.
This study was supported by National Institutes of Health grant NS
The authors thank Allyn W. Kimball, PhD, for statistical consultation; Barbara Cysyk, RN and Madge Morrell, RN, for patient selec-
10 Annals of Neurology Vol 30
No 1 July 1991
tion and clinical monitoring; David Clough, CNMT, for PET scan
acquisition and processing; Robert Smoot, CNMT, for assistance
with radiotracer synthesis; and Julia W. Buchanan and Julie Price for
editorial suggestions.
I. Sherwin A, Quesney F, Gauthier S, et al. Enzyme changes in
the actively spiking areas of human epileptic cerebral cortex.
Neurology 1984;34:927-933
2. Bajorek JG, Lee RJ, Lomas P. Neuropeptides: anticonvulsant
and convulsant mechanisms in epileptic model systems and in
humans. Ado Neurol, 1986;44:489-500
3. Nistico G, ed. Neurotransmitters, seizam and epilepsy I l l . New
York: Raven Press, 1986
4. McNamara JO, Bonhaus DW, Crain BJ, et al. Biochemical and
pharmacological studies of neurotransmitters in the kindling
model. In: Jobe PC, Laird HE 111, eds. Neurotransmitters and
epilepsy. Clifton, NJ: Humana Press, 1987:115-160
5. Goldstein DS, Nadi NS, Scull R, et al. Levels of catechols in
epileptogenic and nonepileptogenic regions of the human brain.
J Nearorhem 1988;50:225-229
6. Pintor M, Mefford IN, Hurter I, et al. Levels of biogenic mines,
their metabolites, and tyrosine hydroxylare activity in the human
epileptic temporal cortex. Synapse 1990;5:152-156
7. Briere R, Sherwin AL, Robitaille Y , et al. Alpha-l-adrenoceptors are decreased in human epileptic foci. Ann Neurol
8. Wyler AR, Nadi NS, Porter RJ. Acetylcholine, GABA, benzodiazepine and glutamate receptors in the temporal lobe of epileptic patients. Neurology 1987;37(suppl 1):103
9. Nadi NS, Wyler AR. A chemical and histochemical architecture
of the human epileptic focus. SOCNeurosci Abstr 1987;13:1077
10. Kreuger H , Eddy NB, Sumwalt M, eds. The phamcologv of
the opium alkaloids, part I . Washington, DC: US Government
Printing Office, 1941
11. Hong JS? Gillin JC, Yang H I T , Costa Ii. Repeated electroconvulsive shocks and the brain content of endorphins. Brain Rts
12. Frenk H. Pro- and anticonvulsant actions of morphine and the
endogenous opioids: involvement and interactions of multiple
opiate and non-opiate systems. Brain Res Rev 1983;6:197-210
13. Tortella FC. Endogenous opioid peptides and epilepsy: quieting
the seizing brain. Trends Pharmacol Sci 1988;9:366-372
14. Holaday JW, Tortella FC, Meyerhoff JL, et al. Electroconvulsive
shock activates endogenous opioid systems: behavioral and biochemical correlates. Ann NY Acad Sri 1986;467:249-255
15. Tortella FC, Long JB. Endogenous anticonvulsant substance in
rat cerebrospinal fluid after a generalized seizure. Science 1985;
2 x 1 106-1108
16. Tonella FC, Long JB. Characterization of opioid peptide-like
anticonvulsant activity in rat. Brain Res 1988;456:139-146
17. Jianguo C, Xueking X. A study on opioid peptides in CSF of
patients with epilepsy. Epilepsy Res 1990;6:141-145
18. Hitzemann RJ, Hitzemann BA, Blatt S , et al. Repeated electroconvulsive shock: sodium dependency and regional distribution
of increased opioid binding sites. Mol P h a m c o l 1987;31:
19. Lee RJ, McCabe RT, Wamsley JK, et al. Opioid receptor alterations in a genetic model of generalized epilepsy. Brain Res
20. Nakata Y , Chang KJ, Mitchell CL, Hon JS. Repeated electroconvulsive shock downregulates the opioid receptors in rat
brain. Brain Res 1985;346:160-163
21. Weiss S, Seeger TF, Ostrowski N, et al. Alterations in opiate
receptor occupancy following arnygdaloid kindling. Soc Neumsci
Abstr 1985;11:1068 (Abstract)
22. Tortella FC. Opioid peptides: possible physiological role as endogenous anticonvulsants. In: Ferrendelli JA, Collins RC, Johnson EM, eds. Neurobiologv of amino acids, peptides and traphic
furtors. New York: Kluwer Academic, 1988:163-179
23. Tortella FC, Robles L, Holaday JW.
The anticonvulsant effects
of DADLE are primarily mediated by activation of delta opioid
receptors: interactions between delta and mu receptor antagonists. Life Sci 1985;37:497-503
24. Frost JJ, Wagner HN Jr, Dannals RF, et al. Imaging opiate
receptors in the human brain by positron tomography. J Comput
Assist Tomogr 1985;9:23 1-2 36
25. Frost JJ, Douglass KH, Mayberg HS, et al. Multicompartmental
analysis of "C-carfentanil binding to opiate receptors in humans
measured by positron emission tomography. J Cereb Blood Flow
Metab 1989;9:398-409
26. Frost JJ, Sadzot B, Mayberg HS, et al. Estimation of receptor
number and affinity for "C-diprenorphine binding to opiate receptors in man by PET. J Cereb Blood Flow Metub 1989;9(suppl
27. Sadzot B, Price JC, Mayberg HS, et al. Quantification of human
opiate receptor concentration and affinity using high and low
specific activity "C-diprenorphine and positron emission tomography. J Cereb Blood Flow Metab 1991;11:204-219
28. Frost JJ, Mayberg HS, Sadzot B, et al. Comparison of "Cdiprenorphne and "C-carfentanil binding to opiate receptors in
man by positron emission tomography. j Cereb Blood Flow Metab
29. Frost JJ, Mayberg HS, Fisher RS, et al. Mu-opiate receptors
measured by positron emission tomography are increased in
temporal lobe epilepsy. Ann Neurol 1988;23:231-237
30. Pfeiffer A, Pasi A, Mehraein P, Herz A. Opiate receptor binding
sites in human brain. Brain Res 1982;248:87-96
31. Cross AJ, Hille C, Slater P. Subtraction autoradiography of
opiate receptor subtypes in human brain. Bruin Ref 1987;418:
32. Goodman RR,Snyder SH, Kuhar MJ, Young WS. Differentiation of delta and mu opiate receptor localization by light microscopic autoradiography. Pmc Natl Acad Sci U S A 1980;77:
33. Goodman U, Snyder SH. Kappa opiate receptors localized by
autoradiography to deep layers of cerebral cortex: relation to
sedative effects. Proc Natl Acad Srz USA 1982;79:5703-5707
34. Bancaud G, Hendriksen D, Rubio-Donnadieu F, et al. Proposal
for revised clinical and electroencephalographic classification of
epileptic seizures. Epilepsia 1981;22:489-501
35. Lesser RP, Fisher RS, Kaplan P. The evaluation of patients with
intractable complex partial seizures. Electroencephulogr Clin Neu~ p h ~ 1989;73:381-388
36. Fox FT,Perlmutter JS, Raichle ME. A stereotactic method of
anatomical localization for positron emission tomography. J
Comput Assist Tomogr 1985;9:141-153
37. Schaltenbrand G, Wahren W. Atlas for stereotaxy of the human
bruin. 2nd ed. Chicago: Year Book Medical, 1977
38. Cidis Meltzer C, Bryan RN, Holcomb HH, et al. Anatomical
localization for PET using MR imaging.J Comput Assist Z'omogr
39. Dannals RF, Ravert HT, Frost JJ, et al. Radiosynthesis of an
opiate receptor binding radiotracer: { 1ICIcarfentanil. Int J Appl
Rad lsot 1985;36:303-306
40. Lever JR, Dannals RF, Wilson AA, et al. Synthesis of carbon-I 1
labeled diprenorphine: a radioligand for positron emission tomographic studies of opiate receptors. Tetruhedron Lett 1987;
41. Phelps ME, Huang SC, Hoffman EJ, et al. Tomographic measurement of local cerebral glucose metabolic rate in humans
with (F-l8)2-fluoro-2-deoxy-D-glucose:
validation of method.
Ann Neurol 1979;6:371-388
42. Engel J Jr, Kuhl DE, Phelps ME, Mazziotta JC. Inrerictal cerebral glucose metabolism in partial epilepsy and its relation to
EEG changes. Ann Neural 1982;12:510-517
43. Theodore WH, Newmark ME, Sat0 S, et al. 18F-Fluorodeoxyglucose positron emission tomography in refractory complex
partial seizures. Ann Neurol 1983;14:429-437
44. Gloor P, Olivier A, Quesney LF, et al. The role of the limbic
system in experimental phenomena of temporal lobe epilepsy.
Ann Neurol1982;12:129-144
45. Seeger TF, Sporzo GA, Pert CB, Pert A. In vivo autoradiography: visualization of stress-induced changes in opiate
receptor occupancy in the rat brain. Brain Res 1984;305:303311
46. Xie CW, Mitchell CL, Hong JS. Perforant path stimulation differentially alters prodynorphin mRNA and proenkephahn
mRNA levels in the entorhinal cortex-hippocampal region.
Brain Res Mol Brain Res 1990;7:199-205
47. Xie CW, Lee PH, Takeuchi K, et al. Single or repeated electroconvulsive shocks alter the levels of prodynorphin and proenkephalin mRNAs in rat brain. Brain Res Mol Brain Res 1989;
48. Theodore WH, Blasberg R, Lerdeman D. PET imaging of opiate
receptor binding in human epilepsy using 18F-cyclofoxy. NeurolOW 1990;40(suppl 1):257
49. Cohen RM, Carson RE, Channing M, et al. F-18-cyclofoxy PET
imaging in man. J NuclMed 1988;29(suppl):796
50. Sherwin AL. Guide to neurochemical analysis of surgical specimens of human brain. Epilepsy Res 1988;2:281-288
51. Savic I, Persson A, Roland P, et al. In vivo demonstration of
reduced benzodiazepine receptor binding in human epileptic
foci. Lancet 1988;2:863-866
Mayberg e t al: Opiate Receptors in Epilepsy
Без категории
Размер файла
1 042 Кб
using, opiate, tomography, emissions, receptors, quantification, positron, nonцmu, temporal, epilepsy, lobel
Пожаловаться на содержимое документа