Differential distribution in rat brain of mu opioid receptor carboxy terminal splice variants MOR-1C-like and MOR-1-like immunoreactivity: Evidence for region-specific processing
Abstract
The present study examined immunohistochemically the regional distribution of the mu opioid receptor splice variant MOR-1C by using a rabbit antisera generated against the C-terminal peptide sequences and compared it with MOR-1. Overall, the distribution of MOR-1C–like immunoreactivity (–LI) differed from MOR-1–LI. Both MOR-1C–LI and MOR-1–LI were prominent in a few central nervous system regions, including the lateral parabrachial nucleus, the periaqueductal gray, and laminae I-II of the spinal trigeminal nuclei and the spinal cord. In the striatum, hippocampal formation, presubiculum and parasubiculum, amygdaloid nuclei, thalamic nuclei, locus coeruleus, and nucleus ambiguous MOR-1–LI predominated, whereas MOR-1C–LI was absent or sparse. Conversely, MOR-1C–LI exceeded MOR-1–LI in the lateral septum, the deep laminae of the spinal cord, and most hypothalamic nuclei such as the median eminence, periventricular, suprachiasmatic, supraoptic, arcuate, paraventricular, ventromedial, and dorsomedial nuclei. Double-labeling studies showed colocalization of the two receptors in neurons of the lateral septum, but not in the median eminence or in the arcuate nucleus, even though both MOR-1 isoforms were expressed. Because both MOR-1 and MOR-1C are derived from the same gene, these differences in regional distribution represent region-specific mRNA processing. The regional distributions reported in this study involve the epitope seen by the combinations of exons 7, 8, and 9. However, if other MOR-1 variants containing exons 7, 8, and 9 exist, the antisera would not distinguish between them and MOR-1C. J. Comp. Neurol. 419:244–256, 2000. © 2000 Wiley-Liss, Inc.
Morphine and most traditional opioid analgesics act through mu opioid receptors (Pasternak, 1993; Reisine and Pasternak, 1996). Early studies suggested two subtypes of mu receptors, with differing pharmacologic actions (Pasternak and Snyder, 1975; Pasternak, 1993). More recent investigations based on traditional pharmacologic and molecular approaches have implicated yet another mu subtype responsible for the actions of morphine-6β-glucuronide (M6G) (Pasternak and Standifer, 1995). A mu receptor was cloned (Wang et al., 1993; Eppler et al., 1993; Chen et al., 1993; Uhl et al., 1994) shortly after the identification of a clone encoding a delta receptor (Kieffer et al., 1992; Evans et al., 1992). Antisense studies quickly established the importance of MOR-1 in morphine analgesia (Rossi et al., 1994), but detailed antisense mapping studies revealed intriguing differences between morphine and M6G, suggesting that they acted through distinct receptors related to MOR-1 (Rossi et al., 1995a,b, 1997). Yet, only a single mu receptor gene has been identified (Min et al., 1994; Giros et al., 1995; Liang et al., 1995; Mayer et al., 1996). Knockout mice also have confirmed the importance of MOR-1 in morphine analgesia (Matthes et al., 1996; Sora et al., 1997; Loh et al., 1998; Schuller et al., 1999).
Alternative splicing has been observed with several G-protein–coupled receptors, including the somatostatin 2 (Vanetti et al., 1998), dopamine D2 (Guiramand et al., 1995), prostaglandin EP3 (Namba et al., 1993), and serotonin receptor subtypes 5-HT4 and 5-HT7 (Lucas and Hen, 1995). Two isoforms of MOR-1 also have been reported (Bare et al., 1994; Zimprich et al., 1995). Both involve splicing at the position of exon 4. In MOR-1A exon 4 is absent. The predicted coding region extends beyond the normal splice site in exon 3 to give an additional four amino acids. MOR-1B contains an alternatively spliced exon 5 instead of the original exon 4. The binding properties of MOR-1B and MOR-1 are similar, but their desensitization properties and regional distributions differ (Schulz et al., 1998).
We recently identified three new splice variants of MOR-1 in the mouse caused by alternative splicing of four additional exons (6, 7, 8, and 9) downstream from the original exons (Pan et al., 1999). In these variants, combinations of the four exons are alternatively spliced to replace the original exon 4 and, thereby, generate the three new variants, all of which contain the same exons 1, 2, and 3 as originally reported in MOR-1 (Fig. 1). Although these variants differ from MOR-1 only at the intracellular tail of the receptor, there are subtle differences in their binding selectivity profiles. In MOR-1C, the 12 amino acids encoded by exon 4 in MOR-1 are replaced by 52 amino acids derived from the combination of the new exons 7, 8, and 9. In binding studies, MOR-1C binds morphine and other traditional mu opioids potently and selectively with affinities similar to MOR-1. However, dynorphin A is significantly more potent against MOR-1C, competing binding in the MOR-1C expressing cells fivefold more potently than in the MOR-1 cells. In the present study, we have examined the regional distribution of MOR-1C–like immunoreactivity (LI) in the rat central nervous system and compared it with that of MOR-1.
Schematic representation of the various MOR-1 splice variants. Schematic of the MOR-1 gene with the additional exons and the recently isolated variants. The epitopes used to generate the antisera are underlined. Please note that the MOR-1 epitope also includes the adjacent three amino acids encoded by exon 3 (not shown). The sequences shown for MOR-1 are identical for both rat and mouse. The MOR-1B sequence is from rat, MOR-1A from human, and MOR-1C, MOR-1D, and MOR-1E from mouse.
MATERIALS AND METHODS
Generation of anti-peptide antisera
Polyclonal antisera against a 20-residue peptide (KSCMDRGMRNLLPDDGPRQE) from a unique predicted amino acid sequence of MOR-1C was generated by Multiple Peptide Systems (San Diego, CA) in New Zealand White rabbits. Five milligrams of the purified peptide was coupled to keyhole limpet hemocyanin (KLH) with glutaraldehyde as the cross-linking agent, in a ratio of 1 part peptide to KHL (w/w). The conjugate was suspended in phosphate buffered saline (PBS) buffer (1 mg/ml, pH 7.4), emulsified by mixing with an equal volume of complete Freund's adjuvant (total volume of 0.6 ml) and injected into three subcutaneous dorsal sites for the primary immunization (day 0). Subsequent immunizations were performed by using incomplete Freund's adjuvant at days 14, 42, and 56. Sera were obtained at days 52, 66, and 70. Immunoaffinity antibody purification was carried out with 3 mg of synthetic purified peptide coupled to 3 ml of agarose gel through the N-terminal amino group. After deactivation of the remaining activated sites, the gel was equilibrated in a physiological buffer of pH 7.5 at room temperature. Five milliliters of crude serum was applied to the peptide bound gel. After washes, the adsorbed protein was eluted with a glycine-HCl buffer at low pH (2.5) and collected in a Tris buffer (pH 8) designed to neutralize the elution buffer. For all sera, the antipeptide antibody titer was determined with an enzyme linked immunosorbent assay (ELISA) with free peptide as coat (100 pmoles/well). Results are expressed by the reciprocal of the serum dilution that results in an OD492 of 0.2 (detection with horseradish peroxidase anti-rabbit immunoglobulin G (IgG) conjugate and peroxidase dye). After affinity purification the titer from rabbit 46494 was 85,200 with a concentration of 1.7 mg/ml. Sera from rabbit 46494 were used in this study.
The MOR-1 antiserum from guinea pig (Chemicon, Temecula, CA) was generated from the terminal 15 amino acids comprising the intracellular carboxy tail, corresponding to the complete exon 4 and the adjacent three amino acids encoded by exon 3. Staining with this antiserum was indistinguishable from a similar antiserum generated in rabbits (Incstar, Stillwater, MN; data not shown).
Immunostaining of transfected Chinese hamster ovary cells with MOR-1C antisera
Antibody specificity was determined by using EcR-CHO cells transfected with receptor murine MOR-1 or MOR-1C. Chinese hamster ovary (CHO) cells (ATCC, Manassas, VA; passage 3) were grown in F12 supplemented with 10% fetal bovine serum. Transient transfections were performed with 1.5 μg of recombinant plasmids for mouse MOR-1C receptor by the DEAE-dextran precipitation procedure onto semiconfluent cells. Three days after transfections, cells were fixed with 4% formaldehyde in 0.1 M phosphate buffer, pH 7.4 (PB). Cells were then washed with PB and incubated for 14 hours with the MOR-1C antiserum diluted at 1:500 (1.4 μg/ml) in PB containing 1% normal goat serum and 0.3 Triton X. After rinsing with PB, cells were incubated for 2 hours with 1:600 dilution of goat anti-rabbit IgG conjugated with the indocarbocyanine Cy-3™ (Jackson ImmunoResearch, West Grove, PA).
Western analysis of MOR-1C
For immunoblot analysis, membranes from control and transfected CHO cells were pelleted in a microcentrifuge and resuspended in sample buffer (pH, 6.8). Briefly, protein samples were run on 12.5% sodium dodecyl sulfate-acrylamide gels and transferred onto nitrocellulose. Nitrocellulose was blocked for 1 hour in 10% nonfat dry milk in PBS and then incubated with the antiserum diluted at 1:1,000 in PBS containing 0.05% Tween 20 for 1 hour at room temperature. Immunoreactive bands were detected with goat anti-rabbit IgG conjugated with horseradish peroxidase (1:10,000) for 1 hour and developed with a chemiluminescent detection system (Renaissance NEN, Boston, MA). Molecular weights were determined by using colored protein molecular weight markers (Rainbow Markers, Amersham, Arlington Heights, IL).
In vivo studies
Experiments were performed on 10 male Sprague-Dawley rats, each weighing 230–270g, and on 20 male CD1 mice, each weighing 20–30g (Charles River Laboratories, Raleigh, VA).
Antisense studies.
An antisense sequence to exon 8 (5′ GGG CCA TCA TCA GGA AGA AGG 3′) corresponding to nucleotides 1446-1466 of the cDNA (GenBank AF074974) and a mismatch (5′ GGG CAC TCA TAC GAG AAG AGG 3′) oligodeoxynucleotide were synthesized by Integrated DNA Technologies (Coraville, IA). Oligodeoxynucleotides (15 μg in 3 μl of 0.9% saline) were given intrathecally (i.t.) under halothane anesthesia between L5 and L6 with a 30-gauge needle daily for 5 days. Another control group received an equal volume of saline. On the sixth day, mice were perfused.
Dorsal rhizotomies.
Surgical procedures were performed under general anesthesia (ketamine 100 mg/kg, i.p.) supplemented with local anesthesia (2% xylocaine). A dorsal hemilaminectomy from vertebrae cervical 4 (C4) to thoracic 2 (T2) was made on the right side, and C4 to T2 dorsal roots were sectioned through small individual openings of the dura mater under a dissection microscope. For each root section, a 1- to 2-mm portion was excised after application of xylocaine to the root. Care was taken to avoid blood vessels that course along the dorsal roots. The wound was washed with saline solution and then overlying muscles and skin were sutured. After recovering from anesthesia, the animals were housed in individual cages. Seven days after the surgical procedure, the rats were perfused.
Immunohistochemistry
After an injection of sodium pentobarbital (100 mg/kg, i.p.), the rats or mice received an intracardiac perfusion of 0.1 M PBS, pH 7.4 (50 ml and 20 ml, respectively) followed by 4% formaldehyde in 0.1 M PB, pH 7.4 (300 ml or 50 ml, respectively). After the perfusion, the brain, spinal cord, and dorsal root ganglia were removed, postfixed for 4 hours in the same fixative and then cryoprotected overnight in 30% sucrose in 0.1 M PBS. Immunostaining was performed on 40-μm sections cut in the coronal plane on a freezing microtome or 15-μm sections cut with a cryostat in the case of dorsal root ganglia. Sections were divided into three sets. One set of sections was reacted with MOR-1, one with MOR-1C, and the last set was used for double labeling with both antibodies.
Immunostaining was performed according to the avidin-biotin peroxidase method of Hsu et al. (1981). Sections were incubated with a solution of 0.1 M PBS with a 3% normal goat serum and 0.3% Triton-X. The blocking solution was removed from the tissue, and the sections were incubated overnight at room temperature in the primary antiserum diluted at 1:500 (1.4 μg/ml). The sections were washed and then incubated in biotinylated goat anti-rabbit IgG (1:200) and avidin-biotin-peroxidase complex (1:100) (Vector Laboratories). To localize the horseradish peroxidase immunoreaction product, we used a nickel-intensified diaminobenzidine protocol with glucose oxidase adapted from Llewellyn-Smith and Minson (1992). Finally, the sections were washed in PB, mounted on gelatin-coated slides, dried, and cover-slipped with DPX (Aldrich, Milwaukee, WI).
For immunofluorescence, the tissue sections were incubated for 30 minutes at room temperature in a blocking solution of 3% normal donkey serum in PBS with 0.3% Triton-X (NDST). The sections were then incubated overnight in the primary antiserum, diluted to 1:500. After the primary antiserum, the sections were washed 3 times in 0.1 M PBS and then incubated in Cy-3™ conjugated donkey anti-rabbit IgG (1:600; Jackson ImmunoResearch, West Grove, PA) for 2 hours at room temperature. For double-labeling studies, the sections were simultaneously incubated with rabbit anti-MOR-1C and guinea pig anti–MOR-1 (1:2,000, Chemicon) antibodies overnight. The following day, sections were incubated in a mixture of Cy-2™ conjugated donkey anti-guinea pig IgG (1:200) and Cy-3™ conjugated donkey anti-rabbit IgG (1:600), both from Jackson ImmunoResearch. To look for targeting of MOR-1C, we double-labeled sections with synaptophysin (1:200, Sigma, St Louis, MO) and MOR-1C. For immunohistochemical controls, the primary antibody was either omitted, replaced by preimmune sera or adsorbed with one of several concentrations of the synthesized peptide.
Quantification of immunoreactivity
To quantify immunoreactivity loss after rhizotomy or antisense treatment, we used a computer-assisted imaging analysis system (Scion Image PC). By using a CCD camera, we first captured an image of the spinal cord. The video image was then converted to a digital image, each with a gray value ranging from 1 to 254. To determine the fraction of the image taken up by labeled structures, we established a threshold, above which pixels were counted (Abbadie et al., 1996). We measured MOR-1C–LI in five sections of the C7 segment from each rat for the rhizotomy experiment and in five sections of L4 segment from each mouse for the antisense experiment. For statistical analysis, we used a one-way analysis of variance to test for effect of treatment. For multiple comparisons, we used the protected least significant difference Fisher's test. The investigator responsible for measuring the density of the MOR-1C labeling was blind to the treatment of the animal.
Confocal images
For double-labeling studies, we examined some sections by confocal microscopy. The confocal images described below were collected with an axiovert 100 M Zeiss confocal microscope by using Zeiss LSM version 2.01 software. To determine whether MOR-1 and MOR-1C were colocalized in cell bodies or processes, images were captured by using a 212 μm pinhole, a 1.6 μm optical slice, and a 63× objective. Montages were created in Photoshop (5.0, Adobe).
RESULTS
Characterization of MOR-1C antibody and labeling
Immunoblot analysis.
The antiserum reacted weakly with a band of approximately 41-kDa and strongly with a band of approximately 60 kDa in CHO cells stably transfected with the mouse MOR-1C plasmid (Fig. 2). Although there is a faint 60-kDa band in the control CHO cells, this band was very variable from experiment to experiment and was not seen in the majority of studies, suggesting that it was not significant. Treating transfected cells with glycosidase (Boehringer-Mannheim, Indianapolis, IN), which removes sugars, lowered the labeled band to just under the 46-kDa marker (data not shown), close to the theoretical size of 48.8 kDa which is based on the predicted protein structure.
Western blot analysis of MOR-1C immunoreactivity. Membranes from control CHO cells and CHO transfected with MOR-1C were isolated and Western blots performed, as described in Materials and Methods section. The faint 60-kDa band in the control CHO cells was very variable and was not seen in the majority of studies, suggesting that it was not significant. The receptor is glycosylated. Removing the sugars with glycosidase lowered the labeled band to just under the 46-kDa marker, close to the theoretical size of 48.8 kDa, which is based on the predicted protein structure.
The immunoreactivity was abolished completely when the staining was performed in the presence of an excess of the peptide antigen or with preimmune serum. The antiserum did not seem to react with nontransfected CHO cells, which do not express the receptor.
Transfected cells.
Transiently transfected CHO cells with cDNA encoding MOR-1C were intensely immunolabeled with the antiserum (Fig. 3A). By contrast, no immunoreactivity was detected in nontransfected CHO cells (Fig. 3B) or in cells transfected with cDNA encoding MOR-1. Immunolabeling was abolished when the incubation was performed with either preimmune serum or with immune serum preabsorbed with the peptide antigen (10 nM) or when the primary antibody was omitted.
Confocal images of CHO cells immunocytochemically stained for MOR-1C. Control CHO cells and cells transiently transfected with cDNA encoding MOR-1C were fixed and stained, as described in Materials and Methods section. The transfected cells exhibited intense cytoplasmic immunoreactivity (A), whereas the nontransfected cells were immunonegative (B). Scale bar = 10 μm (applies to A,B).
Brain tissue controls.
Absorption controls were performed in rat spinal cord tissue by using the peptide used for immunization. Staining was not detectable when using 10−8 M or higher concentrations of the peptide (Fig. 4). Immunolabeling was greatly decreased in the absence of detergent and abolished when the incubation was performed with preimmune serum or when one if the reaction steps (primary or secondary antibody) was omitted.
MOR-1C labeling after peptide absorption or rhizotomy. Photomicrographs of rat cervical spinal cord illustrating MOR-1C–like immunoreactivity (MOR-1C–LI) were taken after absorption with the peptide (10−8 M) used for immunization (B) or rhizotomy (D). Note the absence of staining after preabsorption with the peptide used for immunization (B) compared with an adjacent section with no preabsorption (A). Seven days after dorsal root section of C4-T2 MOR-1C–LI was decreased by 20–30% in laminae I-II of the C7 segment (arrowheads in D) compared with the contralateral side of the rhizotomy (C). Scale bar = 150 μm (applies to A–D).
Antisense treatment.
To examine the specificity of the antibody, mice were treated intrathecally for 5 days with antisense DNA targeting exon 8 as described in the Materials and Methods section. On the sixth day, the spinal cords were obtained and compared with saline and mismatch treated controls. Antisense treatment decreased the MOR-1C–LI by 40 to 50 % in laminae I-II (Fig. 5). MOR-1C–LI in saline or mismatch treated mice did not differ from untreated mice.
Effects of antisense treatment on MOR-1C–like immunoreactivity. Photomicrographs were taken of mouse lumbar spinal cord after antisense treatment, as described in the Materials and Methods section. The antisense oligodeoxynucleotide directed was injected intrathecally in mice for 5 days (15 μg/day). Note the 40–50% decrease in laminae I-II after the antisense treatment (arrowheads). MOR-1C labeling in saline and mismatch treated mice did not differ from untreated mice. Scale bar = 100 μm (applies to all panels).
Distribution of MOR-1C in rat central nervous system
General observations on MOR-1C–LI.
The distribution pattern of MOR-1C–LI was the same by using either immunofluorescence or the avidin-biotin peroxidase method. At low magnification we observed intense, cloudy immunoreactivity which became punctate with higher magnification. We observed a discrete labeling pattern of individual neurons, with MOR-1C–LI on the plasmalemma of both perikarya and dendrites but also in thin beaded processes that are likely to be axons. The pattern of the labeling also varied in intensity from region to region. For example, the immunoreactivity was very intense in the lamina I-II of the spinal cord (+++), intense to moderate in lamina X of the spinal cord (++) and weak in the nucleus accumbens (+). The comparison of MOR-1C to MOR-1–LI distribution patterns has been summarized (Table 1). The distribution of MOR-1–LI has been extensively studied (Arvidsson et al., 1995; Ding et al., 1996). The aim of the present study is to compare the distributions of MOR-1C–LI with MOR-1–LI.
| Structure | MOR-1-LI | MOR-1C-LI |
|---|---|---|
| Cerebral cortex | + | +/− |
| Piriform and entorhinal cortex | + | + |
| Olfactory system | ||
| Anterior olfactory nucleus | ++ | + |
| Olfactory tubercule | + | + |
| Striatum | +++ | +/− |
| Nucleus accumbens | ++ | + |
| Endopiriform nucleus | + | − |
| Globus pallidus | + | − |
| Entopedoncular nucleus | ++ | − |
| Ventral pallidum | + | − |
| Lateral septum | ||
| Dorsal | − | ++ |
| Intermediate | + | +++ |
| Ventral | + | +++ |
| Medial septum | ++ | ++ |
| Septofimbrial nucleus | ++ | |
| Bed nucleus of the stria terminalis | + | ++ |
| Nucleus of the diagonal band | + | + |
| Medial preoptic area | + | ++ |
| Ammon's horn (CA1, CA2, CA3) | ||
| Stratum oriens | + | − |
| Pyramidal cell layer | ++ | − |
| Stratum rediatum | + | − |
| Stratum lacunosum-moleculare | ++ | − |
| Dendate gyrus | ||
| Granular layer | ++ | − |
| Molecular layer | + | − |
| Polymorph layer | + | − |
| Presubiculum | +++ | − |
| Parasubiculum | +++ | − |
| Amygdala | ||
| Central | ++ | − |
| Basolateral | ++ | − |
| Lateral | + | − |
| Medial | + | +/− |
| Cortical | + | − |
| Subthalamic nucleus | − | + |
| Zona incerta | − | +/− |
| Thalamus | ||
| Anterodorsal nucleus | + | − |
| Anteroventral nucleus | − | − |
| Anteromedial nucleus | + | − |
| Mediodorsal nucleus | + | − |
| Centrolateral nucleus | + | − |
| Centromedial nucleus | + | − |
| Laterodorsal nucleus | ||
| Dorsal part | + | − |
| Ventral part | − | − |
| Lateroposterior nucleus | ++ | − |
| Ventrolateral nucleus | + | − |
| Ventroposteromedial nucleus | − | − |
| Ventroposterolateral nucleus | − | − |
| Reuniens nucleus | ++ | − |
| Paratenial nucleus | ++ | − |
| Paraventricular nucleus | ++ | + |
| Gelatinosus nucleus | + | − |
| Rhomboid nucleus | + | − |
| Thalamic gustatory nucleus | ++ | − |
| Parafasicular nucleus | − | − |
| Posterior nuclear group | + | − |
| Posterior intralaminar nucleus | ++ | − |
| Thalamic reticular nucleus | ||
| Rostral part | ++ | + |
| Caudal part | − | + |
| Lateral geniculate nucleus | +++ | − |
| Medial geniculate nucleus | ++ | + |
| Lateral habenula | ++ | + |
| Medial habenula | +++ | ++ |
| Hypothalamus | ||
| Lateral | + | + |
| Anterior | + | + |
| Median eminence | ++ | ++++ |
| Periventricular nucleus | + | +++ |
| Suprachiasmatic nucleus | ++ | +++ |
| Supraoptic nucleus | − | ++ |
| Arcuate nucleus | + | +++ |
| Paraventricular nucleus | + | ++ |
| Ventromedial | + | ++ |
| Dorsomedial | + | ++ |
| Premammillary nuclei | + | + |
| Supramammillary nucleus | ++ | + |
| Peripeduncular nucleus | ++ | + |
| Substantia nigra | + | − |
| Ventral tegmental area | ++ | − |
| Interpeduncular nucleus | +++ | − |
| Superior colliculus | ||
| Deep gray | + | + |
| Superficial gray | ++ | ++ |
| Intermediate gray | − | − |
| Inferior colliculus | ||
| Dorsal cortex | + | + |
| External cortex | + | ++ |
| Central nucleus | − | +/− |
| Periaqueductal gray (rostral) | + | +/− |
| Periaqueductal gray (caudal) | + | ++ |
| Locus coeruleus | +++ | +/− |
| Parabrachial nucleus | ||
| Lateral | +++ | +++ |
| Medial | ++ | + |
| Nucleus of solitari tract | +++ | ++ |
| Raphe nuclei | ||
| Dorsal | ++ | + |
| Median | +++ | +/− |
| Magnus | + | ++ |
| Obscurus | − | + |
| Pallidus | + | + |
| Gracile nucleus | + | +/− |
| Cuneate nucleus | + | +/− |
| External cuneate nucleus | + | +/− |
| Spinal trigeminal nucleus | ||
| Oral | + | +/− |
| Interpolar | + | +/− |
| Caudal | ||
| Laminae I-II | +++ | +++ |
| Other laminae | + | + |
| Area postrema | + | + |
| Ambiguus nucleus | +++ | +/− |
| Cerebellum | − | − |
| Spinal cord | ||
| Laminae I-II | +++ | +++ |
| Laminae III-IV | +/− | + |
| Laminae V-VI | +/− | ++ |
| Lamina X | +/− | +++ |
| Laminae VII-VIII | +/− | + |
| Lamina IX | − | − |
| Onuf's nucleus | + | +++ |
| Intermediolateral nucleus | + | ++ |
| Lateral spinal nucleus | + | ++ |
| Dorsal root ganglia | + | + |
- 1 Intensity of immunoreactivity: +++, high; ++, moderate; +, low; +/−, very low; −, negative.
Telencephalon.
The distribution of MOR-1C–LI varied among the various regions and differed from that of MOR-1. The olfactory system showed weak MOR-1C–LI. In the olfactory tubercle, scattered neurons were found in the granular cell layer, whereas moderate labeling was observed in the external part of the anterior olfactory nucleus. In the cerebral cortex, the piriform cortex had moderate staining. The motor cortex revealed occasional scattered neurons and varicose fibers, but little reactivity in other cortical regions. Little MOR-1C–LI was present in the hippocampal formation, contrasting with the dense MOR-1–LI observed in the dentate granule cell layer and in the pyramidal cell layer of Ammon's horn, and in the presubiculum and parasubiculum.
MOR-1C–LI was very abundant in neuronal cell bodies of the lateral septum, where cell bodies could be identified throughout the rostrocaudal extent (Fig. 6A,B). The most dorsal part near the corpus callosum also contained a little MOR-1C–LI. Double-labeling studies revealed colocalization of MOR-1C–LI and MOR-1–LI in neurons located in dorsal part of the lateral septum (Fig. 9A–C). More diffuse MOR-1C–LI was observed in the medial septum and in the nucleus of the vertical and horizontal limbs of the diagonal band of Broca. In the dorsal part of both the lateral and medial subdivisions of the bed nucleus of the stria terminalis, MOR-1C–LI was found in neurons and varicose fibers (Fig. 6J).
MOR-1C–like immunoreactivity in rat brain. Brain sections were labeled with MOR-1C antiserum, as described in the Materials and Methods section. Immunofluorescence labeling with MOR-1C antisera was abundant in the lateral septum (A,B), the periventricular hypothalamic nucleus (C,D), the medial eminence (E–G; box shown in F is magnified in G), the arcuate nucleus (H,I; box shown in H is magnified in I), and the bed nucleus of the stria terminalis (J). Note the numerous MOR-1C labeled cell bodies (arrowheads in B, D, and I). LV, lateral ventricule; CPu, caudate putamen; 3V, third ventricule. Scale bars = 200 μm in A,E,H; 100 μm in B,C,D,F,I,J; 50 μm in G.
MOR-1C–like immunoreactivity (MOR-1C–LI) in rat brainstem. Brainstem sections were labeled with MOR-1C antiserum, as described in Materials and Methods. Note the intense MOR-1C–LI in the lateral parabrachial nucleus (A,B; box shown in A is magnified in B), and in the inferior olivary complex (C,D; box shown in C is magnified in D). Scale bars = 200 μm in A,C; 100 μm in B,D.
MOR-1C–LI in rat spinal cord. Spinal cord sections were labeled with MOR-1C antiserum, as described in the Materials and Methods section. Immunofluorescence labeling with MOR-1C antisera was particularly abundant in the superficial laminae (A,B,C,H) and around the central canal (A,D,F). In laminae I-II (C) and V (E) MOR-1C–LI was in thin beaded processes. MOR-1C–LI was also distributed in the medial part of laminae V-VI (arrowheads in D) and of laminae VII-VIII (arrowheads in F). In the L5 segment, the medial motoneuron group of Onuf's nucleus was densely positive for MOR-1C–LI (G) as were regions containing parasympathetic preganglionic neurons in the sacral spinal cord (H,I). LSN, lateral spinal nucleus; DH, dorsal horn; CC, central canal; VH, ventral horn. Scale bars = 350 μm in A; 200 μm in B,D,F,G,H; 100 μm in I; 50 μm in C,E.
Comparison of MOR-1C–like immunoreactivity (MOR-1C–LI) and MOR-1–LI in rat brain. Confocal immunohistochemistry by using MOR-1C and MOR-1 antisera was performed as described in Materials and Methods. The left column shows MOR-1C–LI (red staining), the middle column shows MOR-1–LI (green staining), and the right column shows the distribution of both. When MOR-1–LI and MOR-1C–LI are colocalized, the staining appears yellow. In the dorsal part of the lateral septum (A–C), some neurons express both MOR-1 and MOR-1C (arrowheads in C). More ventrally to the area shown, numerous neurons are only labeled for MOR-1C (not shown). In the arcuate nucleus (D–I), MOR-1 and MOR-1C are not colocalized in fibers (F) or in cell bodies (I). In the median eminence (J–L), MOR-1C and MOR-1 are in separate layers: MOR-1–LI is localized in the most external layer, whereas MOR-1C–LI is more internal. Scale bars = 50 μm in A–C; 100 μm in D–F,J–L; 20 μm in G–I.
The striatum revealed a dramatic difference between the two MOR-1 variants. Intense patches of MOR-1–LI were noted in the striatum, corresponding to patches previously observed in autoradiographic studies (Kuhar et al., 1973; Pert et al., 1976; Atweh and Kuhar, 1977a,b,c; Goodman and Pasternak, 1985). In contrast, the striatum contained little MOR-1C–LI, with only scattered neuronal cell bodies and dendrites showing weak MOR-1C–LI (see lateral part of Fig. 6A). Some MOR-1C–LI fibers were detected in the shell and the core of the nucleus accumbens. In the amygdala, the only MOR-1C–LI was some faint staining in the medial amygdaloid nucleus.
Diencephalon.
The hypothalamic regions contained dense MOR-1C–LI. In the external layer of the median eminence, MOR-1C–LI was extremely intense by using either fluorescence (Fig. 6E–G) or the diaminobenzidine reaction. MOR-1C–LI also was intense in the periventricular nucleus (Fig. 6C,D), the suprachiasmatic nucleus, and the arcuate nucleus (Fig. 6H,I). MOR-1C immunolabeling was moderate in the supraoptic nucleus, the paraventricular nucleus, and the ventromedial and dorsomedial nuclei, and weak in the lateral and anterior nuclei. Neuronal cell bodies were observed in the arcuate (Fig. 6I), the paraventricular, the periventricular (Fig. 6D), the suprachiasmatic, and the supraoptic nuclei. Double-labeling studies failed to reveal any colocalization of MOR-1 and MOR-1C in the arcuate nucleus (Fig. 6D–F), neuronal cell bodies (Fig. 6G–I), or in the median eminence (Fig. 5J–L). In the median eminence, MOR-1–LI was localized to the most external layer, whereas MOR-1C–LI was found more internally. The dorsal premamillary nucleus and the supramamillary nucleus contained weak MOR-1C immunoreactivity, but no labeling was seen in the mamillary nucleus. MOR-1C–LI was intense in the medial habenular nucleus, whereas the staining in the lateral portion was more diffuse and less intense.
In contrast to MOR-1–LI, most regions of the thalamus contained no MOR-1C–LI (see Table 1). MOR-1C–LI labeling was weak in the paraventricular nucleus and the dorsal part of the thalamic reticular nucleus, whereas the ventral part was more intensely labeled. The medial geniculate nucleus also contained weak MOR-1C–LI scattered dorsally. No MOR-1C–LI was observed in the ventral part of the medial geniculate nucleus or in the lateral geniculate nucleus.
Mesencephalon.
The pretectal and tegmental regions and the substantia nigra contained no detectable MOR-1C–LI. MOR-1C–LI was moderate in the zonal layer of the superior colliculus and the superficial gray layer, but weaker in the deep gray layer. In the peripeduncular nucleus, MOR-1C–LI was scattered. MOR-1C–LI was moderate in the external cortex of the inferior colliculus, weaker in the dorsal cortex and faint in the central nucleus. MOR-1C–LI was faint in the rostral part of the periaqueductal gray, more moderate in the caudal part and more intense in the ventrolateral area. No MOR-1C–LI neuronal cell bodies were observed. There was weak MOR-1C–LI labeling of the dorsal portion of the dorsal raphe nucleus and a few processes were labeled in the median raphe nucleus.
Pons.
The locus coeruleus also demonstrated differences between the two variants. The locus coeruleus was one of the most intensely labeled regions with MOR-1, whereas we only observed a few axons with MOR-1C–LI. The lateral parabrachial nuclei exhibited very intense MOR-1 and MOR-1C–LI labeling. Dense MOR-1C–LI was distributed in the external and central subnuclei of the lateral parabrachial, as well as in the Kölliker-Fuse nucleus (Fig. 7A,B). Very few MOR-1C–LI neuronal cell bodies could be seen in the lateral parabrachial area. In the medial nuclei, MOR-1C–LI was faint to weak. A few neurons expressing MOR-1C–LI were localized in the pontine raphe nucleus and in the reticulotegmental nucleus of the pons. Numerous MOR-1C–LI axons could be seen apposed to cells in the nucleus of the trapezoid body.
Medulla.
Moderate to intense MOR-1C–LI processes were observed throughout the hypoglossal prepositus nucleus. The medial, commissural, intermediate, and ventral subnuclei of the solitary tract had intense MOR-1C–LI. The area postrema showed very weak MOR-1C. The ambiguus and the paragigantocellular reticular nuclei contained a few MOR-1C–LI fibers. Faint MOR-1C–LI also was observed in the preolivary nucleus and the granular layer of the cochlear nucleus, whereas labeling was moderate in the facial nucleus. The raphe pallidus and the raphe obscurus had weak MOR-1C–LI, but the immunoreactivity was intense in the raphe magnus. Strong MOR-1C–LI was observed in the dorsal part of the inferior olivary complex (Fig. 7C,D). We also saw intense MOR-1C–LI in laminae I-II of the spinal trigeminal nuclei with moderate MOR-1C–LI in the spinal trigeminal tract. MOR-1C–LI was faint in the externate cuneate, cuneate, and gracile nuclei.
Spinal cord.
The pattern of MOR-1C–LI was comparable at the different segmental levels of the spinal cord, although we recorded segmental differences in both the intensity of the labeling and the number of cell bodies. The lumbar and sacral levels were the most intensely labeled compared with the cervical and thoracic segments. The superficial laminae of the spinal cord contained the highest density of MOR-1C–LI throughout the entire rostrocaudal extent of the cord (Fig. 8A–C). Thin beaded processes were located throughout lamina II (Fig. 8C). Scattered small perikarya were occasionally found in inner lamina II. We also observed a dense concentration of MOR-1C–LI processes and occasional cell bodies surrounding the central canal (Fig. 8D,F–I). The neck of the dorsal horn (laminae V and VI) contained two types of processes. The reticular part contained thin beaded processes (Fig. 8B,E). In the most medial portion, extending from the central canal to laminae V–VI, processes were located adjacent to the dorsal column white matter (arrowheads in Fig. 8D). This was particularly marked in the lumbar cord. Similarly, the medial part of the ventral horn contained MOR-1C–LI, with processes penetrating the adjacent white matter (Fig. 8F). Scattered MOR-1C–LI could also be found throughout laminae VII and VIII (Fig. 8F,G). No MOR-1C–LI was seen in motor neurons except the Onuf's nucleus in L5 and L6 segments (Fig. 8G). Very prominent labeling was recorded in both the medial and lateral groups of the Onuf's nucleus.
Seven days after sectioning cervical dorsal roots, MOR-1C–LI in laminae I-II decreased by 20–30% (Fig. 4C,D), indicating that MOR-1C is localized both presynaptically in primary afferent neurons and postsynaptically. The presence of MOR-1C–LI in primary afferents was confirmed directly with studies on the dorsal root ganglia. MOR-1C–LI was present in dorsal root ganglion neurons of all sizes, but it was more prominent in small to medium neurons.
Although dense MOR-1C and MOR-1 immunoreactivity was present in the superficial layers, MOR-1C–LI was more abundant than MOR-1–LI in deeper laminae of the cord (central canal, neck of the dorsal horn, and ventral horn). In laminae I-II, very few processes contained both MOR-1–LI and MOR-1C–LI (Fig. 10). We observed MOR-1C–LI neurons in the lateral spinal nucleus (LSN, Fig. 8B) located in the dorsal part of the dorsolateral white matter all along the spinal cord. Most of the neurons were small with dendrites that radiated in all directions from the cell body. Although both variants were colocalized in most neurons of the LSN, MOR-1C–LI was far more intense than MOR-1–LI.
Comparison of MOR-1C–like immunoreactivity (MOR-1C–LI) and MOR-1–LI in rat spinal cord. Confocal immunohistochemistry by using MOR-1C and MOR-1 antisera was performed as described in the Materials and Methods section. The left column shows MOR-1C–like immunoreactivity (MOR-1C–LI, red staining), the middle column shows MOR-1–LI (green staining) and the right column shows the distribution of both. When MOR-1–LI and MOR-1C–LI are colocalized, the staining appears yellow. A–I: Images are taken from transverse sections in L5. J–O: Images are taken from parasagittal sections in L4. In the superficial laminae of the spinal cord, MOR-1 and MOR-1C–LI appear to be in the same region when observed at low magnification (C); however, they appear not to be colocalized as seen in cross-sections (I) and in parasagittal sections (O). Scale bars = 200 μm in A (applies to A–C,J–L); 50 μm in D (applies to D–F); 20 μm in G (applies to G–I,M–O).
We saw MOR-1C–LI in sympathetic preganglionic neurons in the thoracic and sacral cord. In the intermediolateral cell column, immunoreactive neurons had dendrites the extended medially to the area surrounding the central canal and the central autonomic area. Similarly, the dendritic branches of many labeled neurons in the parasympathetic preganglionic nuclei arborized medially to the central canal and laterally into the lateral white matter (Fig. 8H,I).
DISCUSSION
The recent identification of four novel exons within MOR-1 and three new splice variants (Pan et al., 1999) illustrates the complexity of the mu opioid receptor family. With the two variants reported several years ago, there are now six different MOR-1 variants differing only in the region of exon 4 (Fig. 1). All the variants retain high affinity for morphine and related mu ligands. However, their regional distributions differ, suggesting that their functional role also may be different. Earlier studies looking at MOR-1B observed dense localization in the olfactory system, with little expression elsewhere (Schulz et al., 1998). With 52 amino acids extending beyond exon 3, MOR-1C is far longer than any of the other variants. It also has two putative casein kinase II phosphorylation sites not present in the others, which have functional implications. Binding studies also demonstrated subtle differences between MOR-1C and MOR-1. However, MOR-1C does not seem to correspond to the M6G receptor, because antisense mapping studies based on the novel exons reveal a loss of morphine but not M6G analgesia. Immunohistochemically, we now find a wide-spread, unique regional distribution of MOR-1C–LI within the brain.
The selectivity of the antisera is always a major consideration in the interpretation of immunohistochemistry. In the current studies, the two antisera seemed to be specific and selective for either MOR-1 or MOR-1C. Adsorption of the antiserum with the peptide used for immunization eliminated labeling, indicating a specific interaction. The lack of cross-reactivity of our antisera for MOR-1 confirmed that it recognized only the sequence encoded by the new exons downstream from exon 3. Similarly, the lack of cross-reactivity of the MOR-1 antibody for cells transfected with MOR-1C indicated that it recognized only the sequence encoded by the original exon 4. Because MOR-1 and MOR-1C share the same exons 1, 2, and 3, antisera targeting exon 3 would show cross-reactivity between them. The observed selectivities were reasonable because the MOR-1 antiserum was generated from the amino acids encoded primarily by exon 4. Only three residues encoded by exon 3 were included in the peptide used to immunize the animals. The MOR-1C antisera was generated from amino acids encoded entirely by the new exons. Our ability to detect MOR-1C from mouse, rat, and guinea pig in Western studies suggested a high homology between these species, at least at the amino acid level. The decrease in labeling after antisense treatments also supports the specificity of the labeling. The regional distributions reported in this study involve the epitope seen by the combinations of exons 7, 8, and 9. The only variant containing this sequence to date is MOR-1C. However, if other MOR-1 variants containing exons 7, 8, and 9 exist, the antisera would not distinguish between them and MOR-1C.
Although our MOR-1 antisera results resembled several reports in the literature (Arvidsson et al., 1995; Ding et al., 1996), they differed from those of Mansour and colleagues (Mansour et al., 1995). These differences may reflect the selectivities of the antisera. The antisera used by Arvidsson and coworkers was generated against the same amino acids as our MOR-1 antiserum and similar to that used by (Ding et al., 1996). However, Mansour and colleagues generated their antisera from a longer sequence, including almost 50 additional amino acids encoded by exon 3 (Mansour et al., 1995). Preabsorption of their antisera with the 12 amino acids encoded by exon 4 of MOR-1 only partially blocked the immunohistochemical staining. One possibility is that the antisera used by Mansour might recognize the exon 3 sequence, an exon that is shared by both MOR-1 and MOR-1C. Thus, it is interesting that Mansour and colleagues found intense staining in regions where we observed labeling by our antisera against MOR-1C and little labeling with the MOR-1 antisera. These regions include the hypothalamic nuclei such as the periventricular nucleus, the median eminence, and the deep laminae of the spinal cord. Thus, the presence of multiple MOR-1 variants has introduced a complexity in the interpretation of these results not appreciated previously and illustrates the need to document the exon encoding the epitope recognized by the antisera.
The differences in immunohistochemical labeling between the two MOR-1 antisera also are interesting when compared with the in situ results seen with a probe looking at MOR-1 (Mansour et al., 1994). In this study, the authors observed labeling in deep regions of the spinal cord, including laminae 7, 8, and 10. These same regions are labeled by the Mansour antiserum, which may recognize exon 3 (Mansour et al., 1995), but not by the antiserum more selective for exon 4, as reported in the literature (Ding et al., 1996) and seen in our own studies. Thus, they provide an example of MOR-1 mRNA levels as determined by in situ studies in a region that seems to express MOR-1C–LI and not MOR-1–LI.
By using light microscopy, MOR-1C–LI seemed predominantly as thin punctuate processes that could represent labeling of either dendrites or axons, a distinction that can only be made through electron microscopy. Some MOR-1C–LI colocalized with synaptophysin, indicating that MOR-1C was targeted into axons. Moreover, we observed a 30% decrease in the superficial laminae after dorsal rhizotomy, indicating that MOR-1C is distributed at both presynaptic and postsynaptic sites. We also found labeling of neuronal cell bodies in the lateral septum, arcuate nucleus, hypothalamic periventricular, paraventricular, suprachiasmatic and supraoptic nuclei, lateral parabrachial nucleus, and the substantia gelatinosa of the spinal cord. By using confocal microscopy, MOR-1C appeared in patches associated with the somatic and dendritic plasmalemma.
It is difficult to compare our results with previous autoradiographic studies because radioligands like DAMGO label both MOR-1 and MOR-1C with similar high affinities. Similarly, in situ hybridization studies would only distinguish between the two isoforms when the probes specifically target exon 4. Probes against the first three exons would identify both variants. However, several areas revealed immunolabeling, despite the absence of receptors autoradiographically or by in situ hybridization. Some examples include the median eminence, which had both MOR-1 and MOR-1C labeling, and the trapezoid nucleus, which contained MOR-1C but not MOR-1 immunoreactivity. On the other hand, mu receptor binding has been localized autoradiographically in layers 1, 4, and deeper layer 6 of the cortex (Lewis et al., 1983), regions lacking appreciable MOR-1–LI or MOR-1C–LI. Additional studies are needed to determine whether these differences result from technical differences between the methods or whether they might represent additional variants not recognized by either antisera.
MOR-1C is present in many relay nuclei of nociceptive pathways. Electrophysiological studies have reported opioid effects in neurons in laminae I-II and in laminae V–VI and X (Besson and Chaouch, 1987). Although MOR-1C–LI is found primarily in the superficial laminae of the dorsal horn, staining also was seen in deeper laminae of both the dorsal and ventral horn, including the area surrounding the central canal. MOR-1C–LI also was present in supraspinal regions important in opioid analgesia, including the nucleus raphe magnus, the caudal periaqueductal gray, the external lateral parabrachial nucleus, and the hypothalamus (Basbaum and Fields, 1984; Bernard and Besson, 1990; Cliffer et al., 1991). However, weak MOR-1C–LI was found in the gigantocellular reticular nuclei, in the locus coeruleus where microinjection of opioids produce analgesia (Bodnar et al., 1988) and almost no MOR-1C–LI was observed in thalamic nuclei or in the central nucleus of the amygdala. The spino(trigemino)pontoamygdaloid pathway has been proposed to modulate emotional, behavioral, and autonomic reactions to noxious events (Bernard and Besson, 1990). MOR-1C might also play a role in stress-related functions from its distribution in the nucleus of the solitary tract that regulates gustatory, gastrointestinal, cardiovascular, and respiratory processes. From its large distribution in hypothalamic nuclei, MOR-1C might be implicated in the well-established hormonal and neuroendocrine modulatory actions of opioids.
The regional distributions reported in this study involve the epitope seen by the combinations of exons 7, 8, and 9. The only variant containing this sequence to date is MOR-1C. However, if other MOR-1 variants containing exons 7, 8, and 9 exist, the antisera would not distinguish between them and MOR-1C.
In conclusion, we have found that the epitope present in MOR-1C has a distinct regional distribution that differs from that of MOR-1. The functional significance of these regional differences is not yet clear. It will be interesting to uncover the mechanisms responsible for the region-specific processing of the MOR-1 gene.
Acknowledgements
The National Institute on Drug Abuse gave a Mentored Scientist Award to Y.X.P. and a Senior Scientist Award to G.W.P., and the National Cancer Institute gave a core grant to Memorial Sloan-Kettering Cancer Center.
