Differential brain opioid receptor availability in central... : PAIN (original) (raw)
1 Introduction
Although the role of the endogenous opioid system (EOS) in pain modulation is widely acknowledged, its behaviour in pathological conditions remains very incompletely understood. Activation of EOS by experimental painful stimuli induces an increase of endogenous opioid peptide release leading to enhanced opioid receptors occupancy. This effect has been documented in animals and humans (Albe-Fessard et al., 1985; Iadarola et al., 1988; Zangen et al., 1998; Zubieta et al., 2001; Bencherif et al., 2002), including patients with inflammatory pain in whom decrease of opioid receptor (OR) binding was interpreted as secondary to endogenous opioid release (Jones et al., 1994). In contrast, EOS changes have rarely been documented in chronic neuropathic pain to date. Willoch et al. (2004) recently described reduced opioid receptor binding in 5 patients with central post-stroke pain (CPSP), which was considered to reflect a sustained increase in the release of endogenous opioids. A different interpretation was proposed by Jones et al. (2004), who studied 4 CPSP patients and concluded that OR binding decrease reflected loss of OR-bearing neurons, rather than enhanced endogenous secretion.
Pain-related secretion of opioid peptides is a reactive phenomenon that occurs following chronic and inescapable stress conditions (Vaccarino et al., 1999) including pain, whether central or peripheral (Fields, 2004), while the loss of brain receptors is more likely to occur following lesions of the central nervous system. It can be anticipated that, if OR binding decrease in neuropathic pain (NP) reflects reactive endogenous opioid peptide secretion, this should be observed in all types of NP, whether central or peripheral. Conversely, if this phenomenon is secondary to actual loss of opioid receptors, it should be observed in central NP exclusively. Since no study of OR has yet been conducted in samples comprising both central and peripheral forms of NP, this interpretive ambiguity cannot be solved yet.
In this study, we used the non-selective opioid antagonist [11C]diprenorphine and positron emission tomography (PET) to investigate changes of central opioid binding in a sample of patients with NP comprising a balanced number of central and peripheral forms. By comparing the two groups among them, as well as each group against a sample of healthy controls, we aimed at contrasting the hypotheses of increased secretion of endogenous opioids versus loss or inactivation of receptor-bearing neurons in NP. Since central and peripheral forms of NP have been shown to react differently to opioid therapy (Dellemijn and Vanneste, 1997; Benedetti et al., 1998; Rowbotham et al., 2003), possible differences in opioid binding in central and peripheral NP may provide pathophysiological hints underlying their different sensitivity to opioids.
To allow meaningful inter-hemispheric comparisons of OR binding, patients selected for the study had unilateral or strongly lateralised pain. Also, to ensure reliability and reproducibility of OR binding measures, and to increase the robustness of the data, each subject underwent two PET studies at 2-weeks interval, under identical medication. The two consecutive exams were scrutinized quantitatively to ensure a stable rate of OR binding between them.
2 Patients and methods
2.1 Patients
Fifteen patients (10 men) aged 30–67 years (mean age 51.9 years), all but one right handed, were included in this research. They were studied prospectively between 2001 and 2004, and all of them suffered from chronic NP, which was of supraspinal origin (CPSP) in 8 patients, and of peripheral or radicular origin in 7 (4 proximal nerve injuries, 2 plexus avulsions and 1 combined root/spinal injury). The mean age was, respectively, 55.2 (±11) and 48 (±10) years in the CPSP and peripheral NP groups.
Pain was strictly unilateral in 14 out of the 15 cases (7 right, 7 left), and strongly asymmetrical in the remaining one (CPSP patient no. 7, visual analog scale score 8/10 in left side, 3/10 in right side). Patients were in good general health, excepting motor and sensory symptoms related to their neurological disease. All of them had continuous, paroxysmal and provoked (allodynic) components of pain in the affected territory, with variable degrees of associated sensory and/or motor deficits (see Table 1 for clinical data). Patients were on various combinations of antiepileptics, antidepressants, and/or paracetamol, at comparable doses in both groups, as described in Table 1. For ethical reasons, medication was kept unchanged during the whole inclusion period. Mean ratings of continuous and paroxysmal pain, under medication, were, respectively, 7 and 8.8/10 on a scale where 0 was ‘no pain’ and 10 ‘the worst imaginable pain’. These ratings remained stable during the inclusion period. Pain was considered refractory to medical therapy, as subjective ratings remained superior to 5/10 despite appropriate pharmacological polytherapy. Accordingly, all patients had entered a therapeutic program and they were potential candidates to electrical neurostimulation (spinal or cortical) for pain control. All patients had undergone a trial of conventional opiates (oral morphine titration) with no response and/or severe side effects in 7 cases, and mild effect on pain in 8 (5 out of 7 patients with peripheral NP were in this latter group). All patients were off opioid medication since at least two months before entering the protocol, and remained so during the whole inclusion period of the study.
Clinical characteristics of patients
The neuropathic character of pain had been established on the basis of clinical data, neurological examination, and detailed electrophysiological investigations (Laser and Somatosensory Evoked Potentials, Electromyogram). Lesion location and extent were documented by magnetic resonance imaging (MRI). Anatomical MR acquisition parameters were as follows: TR=1970; TE=4ms; voxel size=1×1×1mm; 120 slices and 256×256 matrix size.
All patients gave their written informed consent to the study, which was conducted according to the declaration of Helsinki (Rickham, 1964), and approved by the Local Medical Ethics Committee (University Hospital, St. Etienne, France).
2.2 Control subjects
PET-scan data from fifteen healthy subjects were used as a control group. Age and sex distribution in this group were matched with those of the patient group. All subjects were pain free, with no neurological, psychiatric and/or painful conditions, and were not under medication or recreational drug. Mean age was 51.3±9.5 years; 5 control subjects were women.
2.3 PET acquisitions and radioligand
PET acquisitions were performed using a Siemens ECAT HR+ scanner, in three-dimensional mode (63 slices, 2.425mm thick each, and a transaxial matrix of 128×128 pixels), with a maximum center field resolution of 4.4×4.41×4.41mm3measured with the NEMA protocol (Brix et al., 1997). The non-selective OR antagonist [11C]diprenorphine was synthesized in situ at the PET center of Lyon (CERMEP), using the methylation method described by Luthra et al. (1994). Radiochemical purity was >95% and specific activity at injection time was in the range of 300–800mCi/μmol.
During the PET study, each subject wore an individually made soft head mould to minimize movements during acquisition time. Using a laser beam on facial landmarks, the head was positioned so as the x_–_y plane of the sensors ring was parallel to the cantho-meatal line. The acquisition room was on dim light, and the subject kept his eyes and ears covered to ensure partial sensory deprivation.
Prior to injection, a 10min transmission scan using an external rotating rod source of 68Ge was performed in order to correct radiation attenuation of tissues of different densities. 4,6 to 8mCi of high specific activity [11C]diprenorphine was injected intravenously, then a 70-min continuous acquisition was performed on a dynamic mode of 37 frames (15×20s, 15×120s, 7×300s).
To assess the reliability of OR binding evaluation, each patient underwent two PET sessions (at two weeks interval), under identical conditions of acquisition and medication. Mean pain scores remained unchanged from one session to the other. The control group underwent only one diprenorphine PET session.
2.4 Data processing
When head movements were detected despite head contention, re-alignment of frames was performed using a home-made software based on the algorithms described by Woods et al. (1998), which normalized each time frame to the frame 18. All frames were decay-corrected.
Data processing was performed using, first, CAPP (Clinical Application Programming Package, CTI, Knoxville) and then SPM 99 (Statistical Parametric Mapping, Welcome department of Cognitive Neurology, London, UK). The first step included orientation of the inter-hemispheric scissure along the y_–_z plane at _x_=0, and realignment of antero-posterior commissural line to overlap the x_–_y plane at _z_=0. Re-orientation parameters were extracted from the mean summated image value and applied to all frames. Parametric images of the Binding Potential (BP) index for [11C]diprenorphine were obtained using the reference tissue model (Frost et al., 1989; Willoch et al., 2004) with occipital cortex as reference ([pixel value−occipital value]/occipital value) applied to all brain voxels of the mean summated value of radioactivity image from frame 20 to 37. The occipital values were calculated using volumes of interest drawn by the investigator, and having the same shape and three-dimensional size for all patients. This was performed by using CAPP (Clinical Application Programming Package, CTI, Knoxville) software. Mean radioactivity curve over the 70min acquisition in the thalamus (high receptor density) showed stable corrected decay values after the frame 19 (Fig. 1); therefore, frames 20–37 were considered as reflecting [11C]diprenorphine binding to receptors with minimal flow effect. For inter-hemispheric comparisons, we standardized the hemispheres across patients as “contralateral” and “ipsilateral” to pain. Hence, parametric images were flipped along the x axis when necessary, following the procedure used by Peyron et al. (1998) and Willoch et al. (2004), so that the side of pain was the right side in all patients (thus the left hemisphere is contralateral to pain in all images).
Comparative time–activity curves (Bq/cc) during 70 min following injection of [11C]diprenorphine, in high (thalamus), intermediate-high (mid-anterior cingulate cortex ACC), intermediate-low (sensori-motor cortex), and very low (occipital lobe) binding regions. Stable values of the curve after time-frame 19 in the thalamus and cingulate reflect ligand fixation to receptors, while the low density of opioid receptors in the occipital lobe results in rapid decline of the activity with time. The ascending part of the curves common to all regions (0–5 min) mainly reflects the blood flow effect after bolus injection.
Individual parametric images were coregistered and, when necessary, resliced with the MR (T1-weighted) of each patient or control subject by application of the translational and rotational parameters of coregistration of the mean summated image (from frame 20 to 37) with the MR. The summated radioactivity image (frame 20 to 37) was used because of more precise anatomic marks for coregistration than individual frames. Each subject's MR was transformed with non-linear basis functions into the standardized space of ICBM (International Consortium for Brain Mapping: www.loni.ucla.edu/ICBM/, McGill university, Montreal, Quebec). Identical transformation parameters as those used for MR were then applied to the previously co-registered parametric images of each subject. These transformations yielded a standardized set of images mutually comparable on a voxel-to-voxel basis. As the two hemispheres of the standardized ICBM space are not strictly symmetrical, MRs flipped along the x-axis were first transformed using non-linear functions into the standardized space of ICBM, and the transformation parameters were then applied to the corresponding flipped parametric images. Using this method, contralateral and ipsilateral hemispheres were “normalized” to the same hemisphere of the standard ICBM space, thus allowing further interhemispheric comparison. Before all statistical analyses, smoothing of images was performed using a 3D-Gaussian Kernel of 10mm to account for residual anatomo-functional variability across subjects. This improves signal to noise ratios and allows inter-hemispheric comparisons on a group basis.
2.5 Statistical analyses
The two PET sessions obtained in each patient at two weeks interval were compared using voxel paired _t_-test with SPM99 software.
Central (_n_=8) and peripheral (_n_=7) groups were compared separately to age- and sex-matched normal volunteers using a between-group one-way ANOVA model with SPM99 software. The side of pain was set to be the right side for the patients, and the corresponding control PET was mirrored when it was the case for the patient's one. Contrasts of interest were (a) “patients (CPSP, peripheral NP) minus control” and (b) “control minus patients”.
For inter-hemispheric (intra-subject) comparisons (hemisphere ipsilateral versus contralateral to pain), each volume unit of one hemisphere was compared with the corresponding volume of the contralateral hemisphere. Since in all image volumes the right hemisphere was ipsilateral to pain (see above), each patient's image was x-flipped, yielding two sets of images: the original set where the right hemisphere was ipsilateral to pain, and the x-flipped set with the right hemisphere contralateral to pain. Then, we compared with SPM the right hemisphere of both sets of images, which was in one case (set 1) contralateral, and in the other (set 2) ipsilateral to pain. This relatively complex procedure was applied since direct interhemispheric comparison is not feasible using SPM software. T-contrasts of interest were (a) “contralateral minus ipsilateral” and (b) “ipsilateral minus contralateral”. Thus, each volume unit of each hemisphere was compared with the corresponding volume of the other hemisphere, on a voxel-wise manner. To account for the effect of random global differences, the global binding index was introduced as a covariate in all comparisons. Interhemispheric comparisons were performed separately for patients with central (_n_=8) and peripheral (_n_=7) neuropathic pain.
Further to interhemispheric comparisons, OR binding in patients with central (_n_=8) and peripheral pain (_n_=7) was compared using a between-group one-way ANOVA model on a voxel-wise manner, with SPM99 software. In the 2 groups, the side of pain was set to be the right side (see above). Contrasts of interest were (a) “CPSP minus peripheral NP” and (b) “peripheral NP minus CPSP”. To account for the effect of random global differences, the global binding index was introduced as a covariate in all comparisons.
For all the above comparisons a number of a priori hypotheses was taken into account, leading to some restriction in the brain volume where statistical analyses were applied. Thus, on the basis of previous opioid receptor imaging using diprenorphine (Sadzot et al., 1990, review in Sprenger et al., 2005), and of our own preliminary data, we constructed an “inclusive mask” of the regions known to contain opioid receptors, using 3D-masks obtained from the Voi Tools module of SPM 99 software (Welcome department of Cognitive Neurology, London, UK). The mask included, in each hemisphere, the insula, thalamus, striatum, caudate nucleus, orbito-frontal, prefrontal and posterior temporal cortices, mid- and anterior cingulate gyrus, and posterior midbrain. For all comparisons the probability threshold for accepting significance at the voxel level was set at p<0.001 (corresponding to _α_=0.05 with Bonferroni correction for multiple resel comparisons), and at p<0.005 at cluster level. An exception was made for the periaqueductal gray region in the posterior midbrain, small area where the cluster level for significance was set at 50 contiguous voxels.
3 Results
3.1 Reproducibility of [11C]diprenorphine distribution
Reproducibility of OR-binding was assessed in the 14 patients who underwent two PET-scan sessions (in the remaining subject data from the second PET-scan were corrupted and unavailable). Voxel-based paired _t_-test analysis showed no significant differences in OR binding between the two PET-scans performed at 2 weeks interval. Also, no significant change in subjective pain scores (24-h average rating of continuous or paroxysmal pain the day of each PET-scan) was disclosed between the two sessions. Therefore, data from both OR measurement were used (as repetitions) for all the following statistical comparisons.
3.2 Opioid binding in central and peripheral NP groups versus control subjects
Relative to control subjects, both peripheral and central pain patients showed significant opioid receptor binding decreases in a number of regions. In the CPSP group binding decrease was clearly asymmetrical (Fig. 2A) and predominated in the hemisphere contralateral to pain in the lateral prefrontal cortex, the insula, medial thalamus, posterior temporal cortex and posterior midbrain. The only region with strictly bilateral BP decrease in CPSP patients was the anterior cingulate, in particular its perigenual region. In contrast, the peripheral NP group showed bilateral and symmetrical decrease in binding which concerned the insula, thalamus (mainly medial), anterior cingulate, posterior temporal and orbitofrontal cortices, posterior midbrain and striatum (Fig. 2B). No region with binding increase was noted in patients relative to controls.
Statistical parametric significance map showing sagittal, coronal and axial projections of structures with reduced opioid binding in patients relative to matched controls: Reduction in the CPSP group was asymmetrical and lateralised to the hemisphere contralateral to pain (A), while in the peripheral NP group (B) binding decrease was symmetrical (see details in text).
3.3 Lateralised changes in OR binding in CPSP patients
Asymmetry in distribution of OR binding was quantified by inter-hemispheric comparisons (see Section 2). In the 8 patients with CPSP, these revealed significant side-to-side differences, with relative decrease of OR binding in a number of regions within the hemisphere contralateral to neuropathic pain. PET-scan results in one representative patient, and statistical (Z) images at the group level, superimposed on a normalized MR, are shown in Figs. 3 and 4. Regions with relative decrease in opioid binding were the lateral prefrontal cortex (mean 32% decrease), posterior temporal cortex (13% decrease), insula (16% decrease), postero-medial thalamus (20% decrease) and posterior part of the midbrain (20% decrease) (Table 2 and Figs. 3 and 4). The midbrain cluster was consistent with the localisation of the periaqueductal gray in the mesencephalon, and was contiguous to the postero-medial thalamic cluster. Table 2 shows the spatial coordinates of the peak of significance within each cluster in the normalized ICBM space, along with the cluster size, _T_-score and significance level. Individual binding potential data are summarized in Table 3.
PET scan showing diprenorphine binding potential (BP) in one representative patient with central pain (patient no. 2): Note reduced BP in the left hemisphere, contralateral to the pain, within lateral prefrontal cortex, insula, posteromedial thalamus and posterior temporal cortex. These areas had also significantly reduced binding potential at the group level in central post-stroke pain patients (see Fig. 4). Although in this patient periaqueductal gray binding was only slightly reduced, this region exhibited significantly decreased BP at the group level.
Clusters of significantly reduced diprenorphine binding potential (BP) contralateral to pain in the group of patients with post-stroke central pain (CPSP), superimposed onto a T1-weighted Magnetic Resonance Imaging (MRI). Both BP and MRI were normalized to the standard stereotactic ICBM space. Stereotactic coordinates (in mm) are relative to the AC-PC line with AC as origin of coordinates. The color scale (bottom right) corresponds to the _T_-score values of displayed clusters. PAG, periaqueductal gray; PTC, posterior temporal cortex; PmT, postero-medial thalamus; LPC, lateral prefrontal cortex.
Interhemispheric comparison in central neuropathic pain group
Individual mean binding potential data, in the two PET-scan sessions, within anatomical regions where a significant BP asymmetry was observed in the CPSP group
All interhemispheric changes in CPSP patients corresponded to relative reductions of OR binding in the hemisphere contralateral to the pain. The mirror contrast (“contralateral minus ipsilateral”) exploring possible relative increases of opioid binding contralateral to pain did not yield any significant result.
To rule out the possibility of asymmetrical regional blood flow decrease leading to changes in the binding potential of diprenorphine, we estimated the regional blood flow (by using data from the first 15 time-frames) in those regions of interest where significant clusters of asymmetrical diprenorphine binding had been found. No significant differences between ipsilateral and contralateral blood flow estimations were observed in the selected regions of interest.
3.4 Changes in OR binding in peripheral NP
Side-to-side inter-hemispheric comparisons in the group with peripheral forms of NP were performed using the same analysis procedures as those used with CPSP patients. No significant interhemispheric decrease or increase in opioid binding could be disclosed in this group.
3.5 Central versus peripheral NP groups
Voxel-wise comparison between CPSP (8 patients) and peripheral (7 patients) NP groups revealed between-group differences in opioid binding restricted exclusively to the hemisphere contralateral to pain. Regions where OR binding was decreased in the CPSP group, relative to peripheral NP, largely converged with those observed in inter-hemispheric comparisons, and comprised the insular cortex (mean 18% decrease), the posterior temporal region (16% decrease) and the lateral prefrontal cortex (30% decrease) (Fig. 5). Table 4 shows the spatial coordinates of the peak of significance within each cluster in the ICBM space, along with the cluster size, _T_-score and significance level. The mirror contrast “CPSP minus peripheral NP” did not show any OR binding increase in the CPSP group relative to the other group.
Regions with significantly reduced diprenorphine binding potential (BP) in the group of patients with central post-stroke pain, relative to patients with peripheral pain. Significant clusters are superimposed onto a T1-weighted MRI, normalized to the standard stereotactic ICBM space. Stereotactic coordinates (in mm) are relative to the AC-PC line, with AC as origin of coordinates. The color scale (bottom right) corresponds to the _T_-score values of displayed clusters. LPC, lateral prefrontal cortex; PTC, posterior temporal cortex; SI, secondary somatic area.
Group comparison of OR binding in CPSP versus peripheral NP
4 Discussion
When each of the patients’ groups was compared with matched control subjects, CPSP patients showed a decrease of the binding potential lateralised to the hemisphere contralateral to pain, while in the peripheral NP group decrease was symmetrical. A within-subject approach based on inter-hemispheric comparison of homologous areas confirmed significant OR binding decrease contralateral to pain in CPSP patients exclusively. Lastly, direct comparison of central versus peripheral patients demonstrated inter-group differences, with OR binding decrease in central patients, exclusively within the hemisphere contralateral to pain.
This pattern of OR changes in our 8 CPSP patients was similar to that reported by Willoch et al. (2004) and Jones et al. (2004) who also described OR binding decrease in brain regions involved in pain processing, mainly contralateral to painful side, in, respectively, 5 and 4 CPSP cases. We believe that this is the first attempt to compare OR binding data from peripheral and central forms of NP, and that our findings lend objective support to a neuro-chemical difference between these two types of NP. Previously reported OR binding decrease in CPSP was interpreted in divergent ways, either as being secondary to loss/inactivation of receptors (Jones et al., 2004), or indicating a reactive increase in the secretion of endogenous opioids (Willoch et al., 2004). Our present findings suggest that, on a background of bilateral endogenous secretion, possibly reactive to pain, there is a lateralised decrease of opioid binding specific to CPSP, probably due to focal loss or inactivation of OR in central pain.
4.1 Opioid binding changes in central NP (CPSP)
The lateralised OR binding decrease in CPSP patients was not restricted to anatomical lesion sites, but rather distributed within structures distant from the lesions. Tissue loss per se could not explain the spatial extent of OR decrease, as lesion sites were spatially restricted in all patients (Fig. 6), and none of the patients’ lesions was constrained to the regions with maximally reduced OR binding (Fig. 4).
Anatomical T1-weighted magnetic resonance images showing the location of brain lesions in the central post-stoke pain group. Coordinates (in mm) are according to the standard stereotactic space of ICBM.
The localisation of binding decrease in CPSP was consistent with previous reports (Jones et al., 2004; Willoch et al., 2004), which also tagged the insula, thalamus, prefrontal cortex and periaqueductal/periventricular gray (PAG/PVG). Bilateral decrease in OR binding was reported by Willoch et al. (2004) within the SII and anterior cingulate cortex (ACC) regions. In our CPSP patients, OR binding decrease in these 2 areas was also bilateral, but reached significance for the ACC only (see Section 3). ACC binding changes, which did not discriminate between our central and peripheral patients, likely pertain to a network of “reactive” binding changes common to patients in the two groups, possibly linked to endogenous secretion of opioids, and different from the lateralised binding decreases that were exclusive of central pain.
The present results in 8 consecutive CPSP patients confirm and extend preceding observations. The relative decrease in OR availability in the hemisphere contralateral to CPSP appears as a robust result that might characterize these patients.
4.2 Findings in peripheral NP
While pain did not differ in intensity, duration, drug intake or any other clinical characteristic, the pattern of OR binding was very different in peripheral NP, in whom no significant lateralised changes in OR binding could be demonstrated. This did not appear to result from changes in intensity, duration of pain or drug intake relative to CPSP patients (see Table 1). The mean VAS in the two groups was perfectly matched for both continuous and paroxysmal pain ratings (Mann–Whitney's _Z_=−0.302 and −1.455; n.s.) and the mean duration of pain was very similar (Mann–Whitney's _Z_=−0.2; n.s.). The number of patients on each medication and corresponding dosages were equivalent in the two groups (see Table 1). Pain had been classified as refractory to pharmacological therapy in all patients, and the intrusive and distressing character of pain was overall comparable between the two groups. Therefore, differences in OR binding between central and peripheral NP may represent a genuine neuro-chemical difference largely independent from any of the demographical or clinical variables studied.
4.3 Significance of lateralised opioid binding decrease in CPSP
The output parameter (BP) used in this study reflects the ratio between the number of receptor sites available for the ligand (_B_max) and the affinity of such receptors for this ligand (_K_D) (Frost et al., 1989). On theoretical grounds, a decrease in BP may result from a reduction in the number of available receptors or an increase in their affinity. Central changes in opioid affinity have never been so far demonstrated in the presence of changes in opioid peptide expression (Millan et al., 1986), and the receptor affinity (_K_D) did not affect the binding of OR antagonists (Yonehara et al., 1983) such as diprenorphine. Therefore, the decrease in opioid binding in our patients is most likely attributable to a reduction in the number and density of available receptors. Among the possible confounding factors influencing receptor number and density, age, gender and analgesic drug intake must be considered. Some of the analgesic drugs taken by our patients, especially antidepressants, may have influenced OR binding (Marchand et al., 2003; Varona et al., 2003; Chen and Lawrence, 2004). Antidepressant dosages were, however, virtually identical in central and peripheral patients (Table 1), and should not have biased the comparisons between them. In particular, the main result of this paper, i.e. selective binding decrease in CPSP contralateral to pain, could not have been explained by medication. Age and gender have been shown to influence μ-opioid receptor binding in humans (Zubieta et al., 1999, 2002). Although women were more represented in the central than the peripheral group, mean age in our patients did not differ significantly between the two groups (T(13)=0.21) and both age and sex in the control subjects were very closely matched to both groups of patients (see Section 2).
Decrease in receptor availability can reflect actual loss of opioid receptors, but may also result from increased endogenous peptide secretion occupying binding sites and rendering them unavailable to the exogenous ligand. Human PET studies showed data compatible with endogenous opioid secretion in both experimental and inflammatory pain (Jones et al., 1994; Zubieta et al., 2001), and this mechanism may have contributed to reduce the available receptors in our chronic pain patients. Indeed, in our patients with peripheral pain the bilateral and symmetrical decrease in diprenorphine binding may have resulted mainly from a reactive endogenous opioid peptide secretion. However, the highly lateralised OR binding decrease observed exclusively in the CPSP group, and not in peripheral pain patients, can hardly be explained by an increase in endogenous opioids. Indeed, endogenous opioid secretion is considered as part of a defensive reaction to inescapable stress conditions, including pain (Vaccarino et al., 1999; Fields, 2004) that depends on pain intensity independent of its aetiology or lesion site, and should therefore have occurred in both peripheral and central pain (as pain intensity and duration were equivalent in both groups). It appears therefore unlikely that such “reactive” opioid secretion may account for the lateralised decrease in opioid binding that was exclusive of CPSP patients.
The lateralised component of opioid binding decrease in CPSP appears most probably related to loss, or metabolic inactivation, of receptor-bearing neurons. While direct neuronal damage could not explain receptor loss, as cerebral structures with decreased OR binding largely exceeded the lesion sites, metabolic diaschisis and transneuronal degeneration in regions interconnected with lesion sites might explain such pattern. Focal lesions can entail metabolic depression at distant but interconnected sites in the same hemispheric network (Feeney and Baron, 1986; Baron et al., 1992), as well as anterograde and retrograde degeneration (Chung et al., 1990). Thus, diaschisis and/or transneuronal changes may have resulted in local metabolic inactivation of opioidergic synapses at regions connected with the damaged areas. Consistent with this hypothesis is the lateralisation of OR binding decrease to the hemisphere where the lesions causing pain were located, as well as the localisation of BP changes in regions interconnected with lesion sites, such as the thalamus, the insula or the PAG. The fact that the lateralised OR binding decrease was associated with central, and not with peripheral NP, also pleads in favour of this hypothesis.
4.4 Clinical relevance
Differences in opioid receptor binding between central and peripheral groups may help to understand their unequal susceptibility to opioid therapy. Better effects of opioids have been documented in peripheral than in central NP conditions (Rowbotham et al., 1991; Dellemijn and Vanneste, 1997). In these latter, and notably in CPSP, opioid effects are inconstant and decrease with time (Dellemijn et al., 1998; Watson and Babul, 1998). Rowbotham et al. (2003) tested the efficacy of oral opioids in 81 NP patients, and reported that peripheral and spinal-related NP were significantly more sensitive to opioids than CPSP (30–39% versus 16%). Also in our series, 5/7 (71%) patients with peripheral NP, but only 3/8 (37%) with CPSP, had partial response to opioids. Our PET results lend support to the notion that loss of opioid sensitivity in CPSP may be due to a decrease in OR central sites available for exogenous opiates.
Acknowledgements
This work was supported by Grants from the “Projet Hospitalier de Recherche Clinique (PHRC)”, the Benoît Foundation, and the Fondation pour la Recherche Médicale (“Equipe FRM”).
References
Albe-Fessard D, Berkley KJ, Kruger L, Ralston HJ 3rd, Willis WD Jr. Diencephalic mechanisms of pain sensation. Brain Res. 1985;356:217-296.
Baron JC, Levasseur M, Mazoyer B, Legault-Demare F, Mauguiere F, Pappata S, Jedynak P, Derome P, Cambier J, Tran-Dinh S, et al. Thalamocortical diaschisis: positron emission tomography in humans. J Neurol Neurosurg Psychiatry. 1992;55:935-942.
Bencherif B, Fuchs PN, Sheth R, Dannals RF, Campbell JN, Frost JJ. Pain activation of human supraspinal opioid pathways as demonstrated by [11C]-carfentanil and positron emission tomography (PET). Pain. 2002;99:589-598.
Benedetti F, Vighetti S, Amanzio M, Casadio C, Oliaro A, Bergamasco B, et al. Dose–response relationship of opioids in nociceptive and neuropathic postoperative pain. Pain. 1998;74:205-211.
Brix G, Zaers J, Adam LE, Bellemann ME, Ostertag H, Trojan H, et al. Performance evaluation of a whole-body PET scanner using the NEMA protocol. National Electrical Manufacturers Association. J Nucl Med. 1997;38:1614-1623.
Chen F, Lawrence AJ. Chronic antidepressant treatment causes a selective reduction of mu-opioid receptor binding and functional coupling to G Proteins in the amygdala of fawn-hooded rats. J Pharmacol Exp Ther. 2004;310:1020-1026.
Chung SK, Cohen RS, Pfaff DW. Transneuronal degeneration in the midbrain central gray following chemical lesions in the ventromedial nucleus: a qualitative and quantitative analysis. Neuroscience. 1990;38:409-426.
Dellemijn PLI, Vanneste JAL. Randomised double blind active-placebo-controlled crossover trial of intravenous fentanyl in neuropathic pain. Lancet. 1997;340:753-758.
Dellemijn PLI, VanDuijn H, Vanneste JAL. Pro1onged treatment with transderma1 fentanyl in neuropathic pain. J Pain Symptom Manage. 1998;16:220-229.
Feeney DM, Baron JC. Diaschisis. Stroke. 1986;17:817-830.
Fields H. State-dependent opioid control of pain. Nat Rev Neurosci. 2004;5:565-575.
Frost JJ, Douglass KH, Mayberg HS, Dannals RF, Links JM, Wilson AA, et al. Multicompartmental analysis of [11C]carfentanil binding to opiate receptors in humans measured by positron emission tomography. J Cereb Blood Flow Metab. 1989;9:398-409.
Iadarola MJ, Brady LS, Draisci G, Naranjo JR. Enhancement of dynorphin gene expression in spinal cord following experimental inflammation: stimulus specificity, behavioral parameters and opioid receptor binding. Pain. 1988;35:313-326.
Jones AKP, Cunningham VJ, Ha-Kawa S, Fujiwara T, Luthra SK, Silva S, et al. Changes in central opioid receptor binding in relation to inflammation and pain in patients with rheumathoid arthritis. Br J Rheumatol. 1994;33:909-916.
Jones AKP, Watabe H, Cunningham VJ, Jones T. Cerebral decreases in opioid receptor binding in patients with central neuropathic pain measured by [11C]diprenorphine binding and PET. Eur J Pain. 2004;8:479-485.
Luthra SK, Brady F, Turton DR, Brown DJ, Dowsett K, Waters SL, et al. Automated radiosyntheses of [6-0-methyl-11C] diprenorphine and [6-0-methyl-11C] buprenorphine from 3-06trityl protected precursors. Appl Radiat Isot. 1994;45:857-873.
Marchand F, Ardid D, Chapuy E, Alloui A, Jourdan D, Eschalier A. Evidence for an involvement of supraspinal delta- and spinal mu-opioid receptors in the antihyperalgesic effect of chronically administered clomipramine in mononeuropathic rats. J Pharmacol Exp Ther. 2003;307:268-274.
Millan MJ, Millan MH, Czlonkowski A, Hollt V, Pilcher CW, Herz A, et al. A model of chronic pain in the rat: response of multiple opioid systems to adjuvant-induced arthritis. J Neurosci. 1986;6:899-906.
Peyron R, Garcia-Larrea L, Gregoire MC, Convers P, Lavenne F, Veyre L, et al. Allodynia after lateral-medullary (Wallenberg) infarct. A PET study. Brain. 1998;121:345-356.
Rickham PP. Human experimentation. Code of ethics of the World medical association. Delaration of Helsinki. Br Med J. 1964;5402:177.
Rowbotham MC, Reisner-Keller LA, Fields HL. Both intravenous lidocaine and morphine reduce the pain of postherpetic neuralgia. Neurology. 1991;41:1024-1028.
Rowbotham MC, Twilling L, Davies PS, Reisner L, Taylor K, Mohr D. Oral opioid therapy for chronic peripheral and central neuropathic pain. N Engl J Med. 2003;348:1223-1232.
Sadzot B, Mayberg HS, Frost JJ. Detection and quantification of opiate receptors in man by positron emission tomography. Potential applications to the study of pain. Neurophysiol Clin. 1990;20:323-334.
Sprenger T, Berthele A, Platzer S, Boecker H, Tolle TR. What to learn from in vivo opioidergic brain imaging? Eur J Pain. 2005;9:117-121.
Vaccarino AL, Olson GA, Olson RD, Kastin AJ. Endogenous opiates: 1998. Peptides. 1999;20:1527-1574.
Varona A, Gil J, Saracibar G, Maza JL, Echevarria E, Irazusta J. Effects of imipramine treatment on delta-opioid receptors of the rat brain cortex and striatum. Arzneimittelforschung. 2003;53:21-25.
Watson CP, Babul N. Efficacy of oxycodone in neuropathic pain: a randomised trial in postherpetic neuralgia. Neurology. 1998;50:1837-1841.
Willoch F, Schindler F, Wester HJ, Empl M, Straube A, Schwaiger M, et al. Central poststroke pain and reduced opioid receptor binding within pain processing circuitries: a [11C]diprenorphine PET study. Pain. 2004;108:213-220.
Woods RP, Grafton ST, Holmes CJ, Cherry SR, Mazziotta JC. Automated image registration: I. General methods and intrasubject, intramodality validation. J Comput Assist Tomogr. 1998;22:139-152.
Yonehara N, Kudo T, Iwatsubo K, Maeda S, Saito K, Inoki R. A possible involvement of the central endorphin system in the autoanalgesia induced by chronic administration of Freund's adjuvant solution in rats. Pain. 1983;17:91-98.
Zangen A, Herzberg U, Vogel Z, Yadid G. Nociceptive stimulus induces release of endogenous beta-endorphin in the rat brain. Neuroscience. 1998;85:659-662.
Zubieta JK, Dannals RF, Frost JJ. Gender and age influences on human brain mu-opioid receptor binding measured by PET. Am J Psychiatry. 1999;156:842-848.
Zubieta JK, Smith YR, Bueller JA, Xu Y, Kilbourn MR, Jewett DM, et al. Regional mu opioid receptor regulation of sensory and affective dimensions of pain. Science. 2001;293:311-315.
Zubieta JK, Smith YR, Bueller JA, Xu Y, Kilbourn MR, Jewett DM, et al. Mu-opioid receptor-mediated antinociceptive responses differ in men and women. J Neurosci. 2002;22:5100-5107.
Keywords:
Positron emission tomography; [11C]Diprenorphine; Opioid receptors; Neuropathic pain; Central post-stroke pain
© 2007 Lippincott Williams & Wilkins, Inc.