| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Studies |
Department of Gynecology and Obstetrics (Y.R.S., M.G.d.C., H.A.Z.) and Department of Radiology, Division of Nuclear Medicine (J.-K.Z., R.F.D., H.T.R., J.J.F.), The Johns Hopkins Medical Institutions, Baltimore, Maryland 21287
Address all correspondence and requests for reprints to: J. James Frost, M.D., Ph.D., Department of Radiology, Division of Nuclear Medicine, The Johns Hopkins Medical Institutions, 600 North Wolfe Street, Nelson B1–130, Baltimore, Maryland 21287. E-mail: jfrost{at}receptor.rad.jhu.edu
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Three main central opioid peptide families have been described in
mammals: the POMC family, which contains ß-endorphin and related
peptides; the proenkephalin family; and the prodynorphin family.
Receptor cloning and pharmacological characterizations have described
three primary types of receptors mediating the effects of endogenous
opioid peptides: µ, 
, and 
(6). Most of the basic research
evidence for interactions with gonadal steroids has involved the POMC
system and the µ-opioid receptors. In the human brain, µ-opioid
receptors have high receptor density and messenger ribonucleic acid
(mRNA) expression in thalamus, basal ganglia, amygdala, and neocortical
and hippocampal cortex, with intermediate densities in the hypothalamus
(7, 8).
Circulating gonadal steroids are also thought to modulate a number of neurotransmitter systems, including the opioid system. In rodents, estrogen receptors in the CNS are primarily located in the hypothalamus, amygdala, lateral septum, and central gray of the midbrain (9, 10). In experimental animals, circulating sex hormones modulate endogenous opioids production and µ-receptor densities in the hypothalamus (11, 12, 13, 14, 15, 16, 17, 18). µ-Opioid binding is also cyclically variable in female rodents, with higher whole brain and hypothalamic densities being observed during low estrogen states (13, 14, 18). It has been proposed that this up-regulation of µ-opioid binding may reflect reduced endogenous opioid release. Hypothalamic POMC mRNA expression, ß-endorphin content, and ß-endorphin-immunoreactive fiber density vary across the estrous cycle in rats, are reduced after ovariectomy, and increase after estradiol and progesterone treatments (19, 20, 21, 22, 23, 24, 25).
Studies exploring interactions between gonadal steroids and the opioid system in nonhuman primates reveal that levels of ß-endorphin in hypophyseal portal blood increase progressively from the mid- to late follicular phase to the luteal phase. If similar changes occur at the level of opioid terminals interacting with GnRH neurons, it could be hypothesized that increased opioid tone (because of either increased receptor levels or increased release of endogenous peptides) may mediate the slowing of LH pulses seen during the luteal phase (2).
Human studies have inferred hypothalamic opioid function based on peripheral plasma ß-endorphin measurements in different conditions and on changes in plasma LH measurements after treatment with opioid antagonists. In humans, a reduction in estradiol as a result of both menopause and oophorectomy appears to be associated with reduced hypothalamic opioid function. This is evidenced by abolishment of the effect of naloxone (a nonselective opioid antagonist) in increasing the release of LH. This effect is reversed by treatment with estradiol and progesterone (26, 27, 28, 29, 30, 31). Whereas these studies present compelling evidence for a link among central opioid activity, gonadal steroids, and GnRH secretion, no one has yet explored these interactions in the living human brain.
Recently, the technique of positron emission tomography (PET) has offered the ability to image and quantitate various receptor types in the living human brain. [11C]Carfentanil, a selective µ-opioid agonist, has been developed and validated for the accurate localization and quantification of CNS µ-opioid receptors (8, 32, 33). In the present work, we sought to study the dynamic alterations in µ-opioid receptors in the human brain during different phases of the menstrual cycle in normal cycling women using PET and to examine the relationship between µ-opioid receptors, and LH pulse secretion and circulating gonadal hormone levels.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Ten women with regular cyclical menses were studied. All subjects provided written informed consent for this clinical investigation, which was approved by the Joint Committee of Clinical Investigation of The Johns Hopkins Medical Institutions. The women [mean age, 25 yr (range, 19–32 yr); mean body mass index, 23 (range, 19–27)] had normal general medical history and physical examinations. Subjects were excluded if they had any of the following: age greater than 35 yr or less than 18 yr; menstrual cycle length greater than 35 days or less than 25 days; hormone use, pregnancy or lactation within the last 6 months; severe premenstrual symptoms; tobacco or illicit drug use; centrally acting medications within the last 6 months; exercise greater than 1 h/day; recent dieting or weight loss; or history of neurological or axis I psychiatric illness in subject or first degree relatives (DSM-IV criteria).
Subjects were studied twice, once in the follicular phase (defined as days 2–8 after the onset of menstrual bleeding) and once in the luteal phase (defined as days 17–23 after the onset of menstrual bleeding). The day of LH surge was determined by commercial urinary LH kits (Clearplan Easy, Whitehall Laboratories, Madison, NJ). The luteal phase testing was performed 3–11 days after the urinary LH surge. No subject was studied less than 2 days before menses. Half the subjects were tested first during the follicular phase, and the other half were tested first in the luteal phase.
For each testing visit, the subject was admitted the evening before to the General Clinical Research Center at the Johns Hopkins Medical Institutions for acclimation to the unit. The subject was given dinner and slept overnight undisturbed. After an overnight fast, an indwelling heparin cannula was placed at 0500 h into a forearm vein. One hour after placement of the cannula, a blood sample was obtained for ß-endorphin measurement. Between 0600–0800 h, two 5-mL blood samples were obtained 1 h apart for estradiol, progesterone, and testosterone measurements. Subjects were served breakfast at approximately 0830 h, and lunch was served at 1100 h. Beginning at 0600 h, 2-mL samples were obtained at 10-min intervals for 9 h for the measurement of LH. All samples were collected in heparinized tubes, refrigerated, centrifuged, and stored at -30 C until analysis.
PET
Immediately after blood sampling was completed the patient underwent a PET scan for quantification and localization of central µ-opioid receptors. All subjects underwent a limited computed tomography scan before the PET studies to align the PET imaging planes within and between subjects, as previously described (34). The imaging plane selected intersected the amygdalae to improve sampling in this relatively small region and was parallel to the long axis of the temporal lobe. PET scanning was then performed in a GE 4096+ scanner (GE Medical Systems, Milwaukee, WI), which acquires 15 simultaneous slices. After positioning and performance of a 10-min 68Ge transmission scan, 19 ± 1 mCi high specific activity (>1000 Ci/mmol) [11C]carfentanil (35) were administered by iv bolus injection. The radiation exposure per scan is estimated as a whole body effective dose equivalent of 0.4 rem and a critical organ (kidney) mean dose equivalent of 1.4 rem. Twenty-five scans of increasing duration (0.5–4 min) were acquired over 90 min. Reconstructed data were then decay and attenuation corrected, and smoothed using a low pass filter to a final resolution of 7.7 mm in-plane at full-width half-maximum, with 6.5 mm separation between planes. PET images were analyzed by averaging image data from 35–82 min posttracer administration and sampling 8 x 8-mm2 regions of interest (ROI) in the averaged images, as previously described (8). ROIs were placed bilaterally in neocortical regions (anterior cingulate, frontal, parietal, temporal, and occipital cortex), amygdala, caudate nucleus, putamen, hypothalamus, and cerebellum, and a single ROI was placed in the pons/midbrain area. Receptor availability measures were then obtained using the ratio (region - occipital cortex)/occipital cortex. The occipital cortex is an area devoid of µ-opioid binding, which is used as a reference region. It provides a combined measure of nonspecifically bound and free tracer concentrations and reflects variables such as tracer metabolization and clearance, and transport of the labeled parent compound across the blood-brain barrier. This method of analysis yields ratio data (without units) proportional to the ratio of receptor density over receptor affinity, as previously described (8) [the ratio Bmax/Kd is also termed binding potential and is the most frequently used measure of receptor binding with PET (36)]. It enables the performance of repeated measures in the same subject, because it does not require more invasive arterial blood sampling. It has also been shown to be largely independent of changes in tracer transport (and therefore variations in regional cerebral blood flow, as tracer transport = flow x extraction) in simulation studies (8).
Assays and LH pulse analysis
Plasma estradiol, progesterone, and testosterone
were determined by solid phase 125I immunoassay using
commercial kits (Diagnostic Products Corp., Los Angeles,
CA). The intra- and interassay coefficients of variance and sensitivity
for each of these RIAs were, respectively: estradiol, 5.3%, 6.4%, and
8 pg/mL; progesterone, 4.7%, 7.9%, and 0.3 ng/mL; and
testosterone, 4.4%, 8.9%, and 4 ng/dL. Plasma ß-endorphin was
determined by a solid phase, two-site immunoradiometric assay using
commercial kits (Nichols Institute Diagnostics, San Juan
Capistrano, CA). The intra- and interassay coefficients of variation
were 4.1% and 9.0%, respectively. The sensitivity was 14 pg/mL.
Cross-reactivity was 100% with human ß-endorphin, 16% with
ß-lipotropin, 0.03% with ACTH, and less than 0.01% with
met-enkephalin, leu-enkephalin, and 
-endorphin. All assays were
performed in duplicate. Samples from all individuals were analyzed in
the same assay.
Plasma LH was determined by a solid phase two-site immunoradiometric assay using commercial kits (Nichols Institute Diagnostics, San Juan Capistrano, CA). This assay has intra- and interassay coefficients of variation of 2.6% and 5.0%, respectively, and a sensitivity of 0.1 mIU/mL (First International Reference Preparation). All samples for an individual were obtained in duplicate within the same assay. LH pulse frequency and amplitude and LH plasma half-life were determined with a standard deconvolution analysis (37, 38).
Because of sample size, LH pulsatility data were examined as follows. As changes in central µ-opioid binding may be caused by factors unrelated to LH pulses, the individual subject’s regional µ binding value was normalized by subtracting the group mean value obtained for both follicular and luteal phase data separately. This eliminates global follicular and global luteal factors that may be contributing to regional receptor variability. Follicular and luteal phases were then combined and correlated to µ-receptor measures for the selected regions with LH pulsatility measures.
Statistics
Intrasubject differences in µ-receptor availability and hormone measures from follicular to luteal phases were examined using two-tailed paired t tests, with a minimum level of statistical significance of P < 0.05. Differences between ovulatory and anovulatory subjects were tested with two-tailed unpaired t tests, with a significance level of P < 0.05. Relationships between circulating levels of gonadal steroids, LH pulsatility, or ß-endorphin, and central µ-opioid receptor measures were examined with Pearson correlations, with P < 0.05 as the level of statistical significance. In some cases, trend correlations (P < 0.1 but > 0.05) are also reported. Given the small number of subjects included in this study, no correction for multiple comparisons was applied. Data are expressed as the mean ± 1 SD.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Mean values for the hormone measurements obtained during
follicular and menstrual phases, as defined above, are shown in Table 1
. In three of the subjects, definite
ovulation (serum progesterone >3 ng/mL) could not be
ascertained despite urinary LH kit data obtained before PET scanning.
Data are presented for all 10 subjects and are divided to discriminate
the ovulatory and anovulatory groups.
|
No significant differences between right- and left-sided
regions were observed (by paired two-tailed t tests).
Therefore, lateralized values were combined for subsequent analyses. No
significant differences in the receptor measure were observed between
follicular and luteal phases when all subjects were included or after
exclusion of the three possible anovulatory subjects (n = 10 and
n = 7; by two-tailed paired t tests, P
> 0.05). Four representative scanning planes are also shown in Fig. 1 during the follicular phase of one
ovulatory subject, with identification of some of the regions of
interest analyzed in this study. As described in prior studies (8),
high µ-opioid binding in vivo was observed in the basal
ganglia (caudate nucleus and putamen), thalamus, and amygdala. Moderate
receptor binding was noted in cortical areas (anterior cingulate
> temporal cortex > frontal cortex > parietal cortex),
cerebellum, and hypothalamus. No specific binding was detected in the
occipital cortex. Values for a representative subcortical region with
high µ-opioid receptor density (amygdala) are shown in Fig. 2.
|
|
). During the follicular phase scan,
regional µ-receptor availability was negatively correlated with
circulating levels of estradiol in the amygdala regardless of whether
only ovulatory or all subjects were included (n = 7 and n =
10; P < 0.01; Fig. 3).
In the hypothalamus, relatively high negative correlations were
observed in both groups, which crossed the significance threshold when
all subjects were included (n = 10; P < 0.05).
None of the other regions examined showed statistically significant
correlations, although moderate negative correlational values were
observed for the thalamus and pons regions.
|
|
). Negative correlations, which did not reach significant levels
during the follicular phase, were now observed with the pons/midbrain
region whether only ovulatory or all subjects were included
(P < 0.05). Circulating levels of estradiol were
significantly and positively correlated with µ binding in the
cerebellum when both ovulatory and all subjects were analyzed
(P < 0.01). Testosterone levels in plasma were also
associated with binding in the cerebellum, but in the opposite
direction (r = -0.84; P = 0.015; Table 3).
|
Central µ-opioid receptors and LH pulsatility
LH pulse data are provided in Table 4. The possible relationship between LH
pulsatility and central µ-opioid availability was examined in three
separate regions: it was examined in the hypothalamus and amygdala, as
these two regions have been previously implicated in the modulation of
GnRH release, and as an internal control region receptor availability
was examined in the thalamus, a subcortical region in physical
proximity to the prior areas, but which is not thought to be involved
in GnRH pulse modulation. We hypothesized that either hypothalamic or
amygdala, but not thalamic, binding would be related to LH pulse
frequency and amplitude.
|
LH pulse number was negatively correlated with µ-receptor
availability in the amygdala (r = -0.60; P =
0.039; Fig. 4). Interestingly,
hypothalamic µ binding was not significantly correlated with the
number of pulses (r = -0.3; P = 0.2). As
expected, thalamic binding was not associated with LH pulse number
(r = -0.05; P = 0.87).
|
), with higher correlational levels in
the left than in the right (left: r = 0.64; P =
0.02; right: r = 0.48; P = 0.1). It did not
correlate with either hypothalamic (r = 0.24; P =
0.47) or thalamic µ-receptor measures (r = 0.24:
P = 0.46). |
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The measures obtained by our methods are proportional to the ratio of Bmax/Kd (binding potential) (36). Changes in µ binding during phases of the rodent menstrual cycle appear to involve changes in receptor density (Bmax), but not receptor affinity (Kd) (39). Therefore, PET scanning measurements are interpreted as reflecting receptor availability and not changes in receptor affinity. As the studies are performed in the living human subject, this technique measures unbound (available for binding) receptors. Any changes observed between experimental states may then reflect changes in receptor occupancy or alterations in actual receptor numbers.
Significant differences in µ-receptor binding between follicular and luteal phases of the menstrual cycle were not observed for any of the brain regions examined. This is in contrast to data obtained in experimental animals, in which modulation of endogenous opioid release, mRNA density, and µ-receptor densities had been noted in vitro in hypothalamic regions (11, 12, 13, 14, 15, 16, 17, 18, 20, 21, 22, 23). As the exact day of the menstrual cycle in which scans were performed could not be standardized in this study because of cycle variability and scheduling issues, there was a range in hormone profiles, and hence, menstrual cycle phase may have been too crude a measure. It is also possible that the changes in available µ binding were small and within the intrasubject experimental variability. In addition, changes within small hypothalamic nuclei may not be detected when the entire hypothalamus is sampled. Interestingly, when the receptor binding data were compared with circulating levels of gonadal steroids separately for each phase of the cycle, several significant correlations emerged.
During the follicular phase, circulating estradiol was negatively correlated with amygdala and hypothalamic binding. This finding was consistent whether including all subjects or only those in whom ovulation was documented. This is consistent with the hypothesis that increasing levels of estradiol stimulate central opioid release, increasing receptor occupancy and hence decreasing available µ-opioid binding, as observed in vivo by PET. In the rodent and sheep brain, the hypothalamus and amygdala are the areas richest in estrogen receptors (9, 40, 41, 42, 43). Both of these regions have been implicated in affecting GnRH pulsatility and reproductive function. Prior in vivo studies in humans have demonstrated increased LH release after naloxone administration, but have not been able to determine the location of this effect (26, 27, 28, 29, 30, 31). Indeed, recent animal work has questioned whether GnRH neurons in fact contain opioid receptors, possibly implying that opioids elicit their response through interneurons (4).
The effect of estradiol on µ-receptor availability in the amygdala and hypothalamus was not seen in the luteal phase. Interpretation of the luteal phase data was complicated by the fact that despite careful subject selection, 3 of 10 subjects did not have definite ovulatory cycles. It is possible that progesterone may have opposing effects from those of estrogen, accounting for this lack of observable effects, as has been suggested by other researchers (44) when examining LH pulsatility. Significant correlations, in opposite directions, between µ-receptor availability in the pons/midbrain region and the cerebellum and circulating levels of estradiol were obtained in the luteal phase. The physiological significance of these observations is unclear at this time. Sex differences in cerebellar glucose metabolism (45) and in µ-opioid receptor binding (33) in humans have been described previously. In this brain area, very low densities of estrogen receptors have been observed in the rodent brain compared to those in other cortical and subcortical regions (42). In regard to the pons/midbrain region, the raphe nuclei, locus coeruleous, ventral tegmental area, and periacqueductal gray all contain high µ-opioid receptor densities. In the latter region, high densities of estrogen receptors have also been described in the rodent and sheep brain (41, 43).
In the luteal phase there were no significant correlations between progesterone and µ binding. This was unexpected, as the role of progesterone in modulating GnRH secretion is well established (46). However, high correlational values were noted for progesterone and µ-opioid binding in several cortical areas (frontal, parietal, and temporal) and in the pons/midbrain region. The small group of ovulatory subjects for the luteal phase analysis precludes making any strong conclusions concerning the relationship between progesterone and µ-opioid binding, which will require examination in future studies.
When LH pulsatility data and documented ovulation for all subjects were
pooled to increase statistical power, LH pulse amplitude and number
were both correlated with µ-opioid receptor binding only in the
amygdala and not in the hypothalamus. Our original hypothesis was that
correlations with LH pulse amplitude would be negative, because
increased opioid tone (and hypothetically less observable in
vivo receptor availability) in the luteal phase has been
associated with increased amplitude of LH pulses (2). We observed
positive correlations between available µ-opioid binding in the
amygdala and LH pulse amplitude. There are two possible interpretations
for this finding: 1) with the increase in opioid receptors, each pulse
of endogenous opioid released is more capable of inhibiting GnRH pulse
amplitude by virtue of its interaction with more available sites; or 2)
the µ-receptors are inhibiting interneurons that inhibit GnRH pulses
(e.g. there is inhibition of an inhibitory system, and
therefore activation). Examples of similar interactions with
intermediary neurons exist for other brain regions. In the ventral
tegmental area, µ-receptor activation stimulates dopaminergic neurons
by inhibiting 
-aminobutyric acid-ergic interneurons (47). The latter
possibility is supported by the recent observation that GnRH-containing
cells do not synthesize opioid receptor mRNA (4).
For LH pulse number the correlation with in vivo amygdala binding was negative. It would indeed be reasonable to expect that LH pulse number would correlate in the opposite direction of LH pulse amplitude, because in the luteal phase an increase in amplitude is associated with decreased pulse frequency. However, central changes in µ-opioid receptors may not completely explain the inverse relationship between LH pulse frequency and amplitude, as at least part of this is likely to be mediated at the pituitary level, based on studies in GnRH-deficient men (48).
An additional finding was that peripheral ß-endorphin levels did not correlate with central µ-opioid binding in any brain region. Likewise, in the primate model, changes seen in hypophyseal portal blood ß-endorphin measurements could not be demonstrated with plasma measurements (2). Furthermore, although peripherally administered ß-endorphin enters the cerebrospinal fluid, it may not enter the brain to any significant degree (49). This emphasizes that peripheral opioid measures do not reflect CNS activity and that central measurements of opioid function are of importance in the study of opioid influence on the menstrual cycle.
In summary, follicular estradiol levels correlated negatively with µ-opioid binding in the hypothalamus and amygdala, presumably reflecting increased central opioid tone and increased receptor occupancy. In addition, LH pulse number and amplitude correlated (in negative and positive directions, respectively) with µ-opioid binding in the amygdala, but not the hypothalamus, suggesting that this brain region is important for the modulation of reproductive function. Interestingly, the amygdala has been frequently associated with emotional responses and the processing of emotional memory (50, 51), providing a possible neurochemical/neuroanatomical link between reproductive and emotional function.
|
Footnotes |
|---|
2 Present address: Department of Obstetrics and Gynecology,
University of Michigan, Room L4224, Women’s Hospital, 1500 East
Medical Center Drive, Ann Arbor, Michigan 48109-0276.
3 Present address: Mental Health Research Institute, University of
Michigan, Neuroscience Building, 1103 East Huron Street, Ann Arbor,
Michigan 48104-1687.
Received February 12, 1998.
Revised September 1, 1998.
Accepted September 14, 1998.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  G. Henriksen and F. Willoch Imaging of opioid receptors in the central nervous system Brain, May 1, 2008; 131(5): 1171 - 1196. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. E. Kennedy, R. A. Koeppe, E. A. Young, and J.-K. Zubieta Dysregulation of endogenous opioid emotion regulation circuitry in major depression in women. Arch Gen Psychiatry, November 1, 2006; 63(11): 1199 - 1208. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. R. Smith, C. S. Stohler, T. E. Nichols, J. A. Bueller, R. A. Koeppe, and J.-K. Zubieta Pronociceptive and Antinociceptive Effects of Estradiol through Endogenous Opioid Neurotransmission in Women J. Neurosci., May 24, 2006; 26(21): 5777 - 5785. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. Benedetti, H. S. Mayberg, T. D. Wager, C. S. Stohler, and J.-K. Zubieta Neurobiological Mechanisms of the Placebo Effect J. Neurosci., November 9, 2005; 25(45): 10390 - 10402. [Full Text] [PDF]
|
|
|
|
|
|
J.-K. Zubieta, J. A. Bueller, L. R. Jackson, D. J. Scott, Y. Xu, R. A. Koeppe, T. E. Nichols, and C. S. Stohler Placebo Effects Mediated by Endogenous Opioid Activity on {micro}-Opioid Receptors J. Neurosci., August 24, 2005; 25(34): 7754 - 7762. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J.-K. Zubieta, T. A. Ketter, J. A. Bueller, Y. Xu, M. R. Kilbourn, E. A. Young, and R. A. Koeppe Regulation of Human Affective Responses by Anterior Cingulate and Limbic {micro}-Opioid Neurotransmission Arch Gen Psychiatry, November 1, 2003; 60(11): 1145 - 1153. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J.-K. Zubieta, Y. R. Smith, J. A. Bueller, Y. Xu, M. R. Kilbourn, D. M. Jewett, C. R. Meyer, R. A. Koeppe, and C. S. Stohler {micro}-Opioid Receptor-Mediated Antinociceptive Responses Differ in Men and Women J. Neurosci., June 15, 2002; 22(12): 5100 - 5107. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
