Abstract
The occurrence of endogenous opioids and their receptors in rat achilles tendon was analyzed by immunohistochemistry (IHC), radioimmunoassay (RIA), and in vitro binding assays. The investigation focused on four enkephalins, dynorphin B, and nociceptin/orphanin FQ. Nerve fibers immunoreactive to all enkephalins (Met-enkephalin, Leu-enkephalin, Met-enkephalin-Arg-Gly-Lys, Met-enkephalin-Arg-Phe) were consistently found in the loose connective tissue and the paratenon, whereas dynorphin B and nociceptin/orphanin FQ could not be detected. The majority of enkephalin-positive nerve fibers exhibited varicosities predominantly seen in blood vessel walls. Measurable levels of Met-enkephalin-Arg-Phe and nociceptin/orphanin FQ were found in tendon tissue using RIA, whereas dynorphin B could not be detected. In addition to the endogenous opioids identified, δ-opioid receptors on nerve fibers were also detected by IHC. Binding assays to characterize the opioid binding sites showed that they were specific and saturable for [3H]-naloxone (Kd 7.01 ± 0.98 nM; Bmax 23.52 ± 2.23 fmol/mg protein). Our study demonstrates the occurrence of an opioid system in rat achilles tendon, which may be assumed to be present also in other connective tissues of the locomotor apparatus. This system may prove to be a useful target for pharmacological therapy in painful and inflammatory conditions by new drugs acting selectively in the periphery.
Keywords
A
Over the past decade, there have been a number of reports on the innervation of musculoskeletal tissues according to specific transmitters, representing the sensory (Grönblad et al. 1985; Hukkanen et al. 1992; Ahmed et al. 1995a; Ackermann et al. 1999) and autonomic nervous systems (Hill and Elde 1991; Ahmed et al. 1995b; Ackermann et al. 2001). Some of these so-called neuropeptides have been shown to promote both nociception and inflammation (Lembeck and Holzer 1979; Lotz et al. 1988; Hart et al. 1995; Snijdelaar et al. 2000).
There also appears to exist a peripheral anti-nociceptive system counteracting the peripheral sensory system. Thus, it was recently reported that intra-articular morphine in conjunction with arthroscopy had a significant analgesic effect in a dose-response manner (Likar et al. 1999). This suggests that opioid receptors are not confined to the CNS but also occur in the periphery. Indeed, there are studies showing opioid receptors on peripheral sensory nerve terminals of the skin (Stein et al. 1990; Coggeshall et al. 1997; Wenk and Honda 1999). Thus far, however, opioid receptors in musculoskeletal tissues have not been identified. In this study we analyzed the occurrence of opioid receptors and endogenous opioids in rat achilles tendon.
Materials and Methods
The study included 77 male Sprague-Dawley (SD) rats (180–200 g), housed five per cage at 21C in a 12:12-hr light:dark cycle with water and pellets ad lib according to the Karolinska Institute protocol. The experiments were approved (no. N 58/98) by the Stockholm North Animal Ethical board.
Radioimmunoassay
Twelve rats were anesthetized by injection of sodium pentobarbital (60 mg/kg IP) and sacrificed by decapitation. The achilles tendons were dissected bilaterally. The samples from the left and right side were pooled to one sample. All dissected tissues were weighed and immediately frozen on dry ice and kept at —70C until neuropeptide extraction. Frozen tendons were boiled for 10 min in 2 M acetic acid, homogenized in a Polytron (15 sec), sonicated (30 sec), and centrifuged at 3000 × g for 15 min. The supernatants were lyophilized, diluted in 2 ml 0.05 M phosphate buffer, pH 7,4, and kept at —20C until analysis (Ahmed et al. 1994). The concentrations of Met-enkephalin-Arg-Phe (MEAP), nociceptin/orphanin FQ (NOC), and dynorphin B (DYN B) were expressed as pmol/g wet weight tissue.
The tissue extracts were subjected to a purification step using an ion exchange procedure before RIA. The method is routinely used to analyze opioid peptides in various tissue extracts and is described in detail elsewhere (Christensson-Nylander et al. 1985; Ploj et al. 2000). The tissue extract was poured onto a small cation exchange column containing SP Sephadex C-25 gel (Pharmacia Diagnostics; Uppsala, Sweden). The peptides were eluted in separate fractions by stepwise elution using a series of buffers containing mixtures of pyridine and formic acid of increasing ionic strength. MEAP elutes in Buffer III (0.35 M pyridine:0.35 M formic acid). N/OFQ and DYN B elute in Buffer V (1.6 M pyridine:1.6 M formic acid). The fractions were dried in a vacuum centrifuge and stored at —20C until peptide analysis.
The peptides were analyzed using specific RIAs for DYN B, MEAP, and N/OFQ, respectively, described in detail elsewhere (Ploj et al. 2000). Before the MEAP analysis was conducted, the samples were oxidized (Ploj et al. 2000). The samples were dissolved in methanol:0.1 M hydrochloric acid (1:1); 25 μl was mixed with 100 μl of antiserum and 100 μl of tracer peptide. The labeled peptide and antiserum were diluted in gelatin buffer. The samples were incubated for 24 hr at 4C. To separate free and antibody-bound peptides in the DYN B and N/OFQ assays, 100 μl of sheep anti-rabbit antiserum (Pharmacia Decanting Suspension; Pharmacia Diagnostics) was added to the samples, which were then incubated for 1 hr before being centrifuged for 10 min at 12,000 × g. The supernatant was discarded and the radioactivity in the remaining pellet was counted in a gamma counter. In the MEAP assay, a charcoal suspension (200 μl), consisting of 250 mg charcoal and 25 mg dextran T-70 in 100 ml 0.05 M sodium phosphate buffer, was added to the samples. After a 10-min incubation period, followed by centrifugation for 1 min at 12,000 × g, the radioactivity in a 300-μl aliquot of the supernatant was counted. DYN B, Tyr14-N/OFQ, and MEAP were labeled using chloramin-T and purified by HPLC. The DYN B antiserum (113B) was generated in rabbits (Christensson-Nylander et al. 1985). It was used at a final dilution of 1:562,000. The antiserum did not show crossreactivity with DYN A(1–17), DYN A(1–8), or other opioid peptides. DYN B 29 crossreacted 1% and “big DYN” (DYN 32) 100%. The detection limit in the assay was 1 fmol/25 μl sample. The N/OFQ antiserum was generated in rabbits (Ploj et al. 2000). In the present study, antiserum 96:2+ was used in a final dilution of 1:112,500. Cross-reactivity with N/OFQ(1–13) was 0.5%; crossreactivity with nocistatin and with the opioid peptides DYN A(1–17), DYN B, DYN A(1–6), DYN 32, DYN B 29, Met-ENK (ME), Leu-ENK (LE), MEAP, and β-endorphin was less than 0.1%. The MEAP antiserum was generated in rabbits (Folkesson et al. 1988); here antiserum 90:3D[II] was used at a final dilution of 1:180,000. Crossreactivity with ME, Met-ENK-Arg6, Met-ENK-Arg6Gly7Leu8 (MEAGL), LE, and DYN A(1–6) was less than 0.1%. The detection limit was 2 fmol/25 μl sample.
Immunohistochemistry
Five rats were anesthetized by injection of sodium pentobarbital (60 mg/kg IP). Intra-arterial perfusion with PBS was performed, followed by perfusion with Zamboni's fixative consisting of 4% paraformaldehyde in 0.2 mol/liter Sorensen phosphate buffer, pH 7.3, containing 0.2% picric acid. The achilles tendons were dissected and immersed in the same fixative for 2 hr at room temperature (RT). All specimens were soaked for at least 2 days in 20% sucrose in 0.1 mol/liter Sorensen phosphate buffer, pH 7.2, containing sodium azide and bacitracin (Sigma Chemicals; St Louis, MO). The tissues were sectioned at 15 μm on a Leitz cryostat and frozen sections were mounted directly on Super-Frost/Plus glass slides and immunostained using the avidinbiotin system. The sections were rinsed two times for 5 min each in PBS. Incubation of the sections with 10% normal goat serum in PBS for 30 min blocked nonspecific binding. Subsequently, sections were incubated with primary antisera for ME, LE, MEAP, MEAGL (all 1:20,000; Peninsula Laboratories, Belmont, CA); NOC, DYN B, δ-opioid receptor (DOR), K-opioid receptor (KOR), μ-opioid receptor (MOR) (all 1:10,000; Peninsula Laboratories), and protein gene product 9.5 (PGP 9.5), a general nerve marker (1:10,000; Ultraclone, Cambridge, UK) overnight in a humid atmosphere at RT. After incubation with the primary antisera, the sections were rinsed in PBS (twice for 5 min) and then incubated with biotinylated goat anti-rabbit antibodies (1:250; Vector Laboratories, Burlingame CA) for 40 min at RT. Finally, the sections were incubated for 40 min with fluorescein isothiocyanate (FITC)-conjugated avidin (1:500; Vector Laboratories). For double staining, after completing the staining steps for the first neurotransmitter, the sections were incubated with avidin blocking solution followed by biotin blocking solution (15 min each). The staining for the receptor was repeated as described for the neurotransmitter, but with a different flourochrome, CY3 (1:5000; Amersham International, Poole, UK). To demonstrate specificity of staining, the following controls were included: (a) pre-adsorption of the primary antisera with excess of homologous antigen (50 μg/ml ME, LE, MEAP, MEAGL; Peninsula Laboratories) for 12 hr at RT; (b) omission of either the primary antiserum, the secondary antibody, or the secondary biotinylated antibody. A Nikon epifluorescence microscope (Eclipse E800; Yokohama, Japan) was used to analyze the sections. T-Max black-and-white and EPL 400 color films (Kodak; Rochester, NY) were used for photography.
Radioligand Binding Assay
The achilles tendons were dissected bilaterally from 60 rats. The tendons were homogenized in 5 v/w of ice-cold 50 mM Tris-HCl buffer, pH 7.4 (Polytron; twice for 30 sec) and sonicated (15 sec). After centrifugation at 40,000 × g for 20 min at 4C, the pellets were resuspended in 30 v/w fresh Tris-HCl buffer and incubated at 37C for 30 min to facilitate degradation and dissociation of endogenous opioid peptides. The centrifugation step was repeated. The final pellet was resuspended in 5 v/w of ice-cold 50 mM Tris-HCl buffer containing 0.32 M sucrose, pH 7.4, and kept at —70C until use.
Aliquots of membrane homogenates (0.6–0.7 mg/ml protein) were incubated with [3H]-naloxone (54.6 Ci/mmol; New England Nuclear, Boston, MA) for 30 min at 25C in a final volume of 250 μl. Reactions were terminated by rapid filtration through Whatman GF/B glass fiber filters pre-treated with 0.1% polyethyleneimine (PEI) using a Brandel Cell Harvester, followed by three washings with 5 ml of ice-cold Tris-HCl buffer, pH 7.4. The bound radioactivity was measured in 4 ml of Ultima Gold scintillation cocktail (Packard) using a Packard 1900 CA liquid scintillation counter. All experiments were carried out in duplicate. Nonspecific binding was measured in the presence of 10 μM unlabeled naloxone.
The protein concentration was determined by the method of Lowry et al. (1951). Scatchard analysis was performed by first-order non-linear regression analysis of the data from saturation isotherms. The binding parameters dissociation constant (Kd), receptor density (Bmax), and Hill coefficient (nH) were calculated from these plots (Spetea et al. 1998). The values presented are the mean ± SEM of three or four independent experiments.
Results
Quantification of Opioids
The RIA focused on MEAP, N/OFQ, and DYN B generated from three different prohormones; proenkephalin, prodynorphin, and pronociceptin. Over all, the measured concentrations were low, some even below the detection limit. Measurable concentrations of MEAP were obtained in 7/11 samples (61%) and of N/OFQ in 4/12 (33%). These proportions of measurable concentrations were significant within the 95% confidence interval, whereas that of DYN B (1/12 = 8%) was not (Table 1). Although MEAP and N/OFQ exhibited a significant proportion of measurable concentrations, only MEAP (0.07 pmol/g) displayed a median concentration above the detection limit (Table 1).
Concentrations (pmol/g) of MEAP, NOC, and DYN B in the achilles tendon of the rat
An out of range value which is not included in the calculations of descriptive statistics.
The concentration level was not detectable.
The concentration level was not determined.
Distribution of Enkephalins
IHC showed that all enkephalins tested, i.e., LE, ME, MEAP, and MEAGL, were present in nerves of tendon tissue (Figure 1). However, no nerve fibers immunoreactive for N/OFQ and DYN B could be detected. LE appeared to be the most abundant opioid compared to ME, MEAP, and MEAGLE in decreasing order. Although they differed in abundance, the tissue distribution displayed a similar pattern. Thus, all four enkephalins predominantly occurred in the loose connective tissue, the paratenon and musculotendinous junction, whereas no enkephalins were found in the tendon tissue proper. In the loose connective tissue surrounding the tendon, the enkephalins appeared as small varicosities around the walls of both large and small blood vessels (Figure 1). In the paratenon and the musculotendinous junction the enkephalins mostly occurred as varicosities in free nerve terminals, without any relationship to blood vessels. Notably, they were more abundant proximally at the musculotendinous junction of the achilles tendon than distally at the bony insertion. The neuronal character of the immunoreactivity was confirmed by PGP 9.5-positive staining (not shown).
Distribution of Opioid Receptors
Of the three opioid receptors (DOR, KOR, MOR) studied, only DOR could be detected by IHC (Figure 2). Positive staining for DOR was found in the loose connective tissue, the paratenon, and the musculotendinous junction. In the loose connective tissue, DOR was predominantly located in the blood vessel walls as small varicosities in nerve fibers penetrating the walls (Figure 2A). In the paratenon and the musculotendinous junction, DOR positivity mostly occurred as varicosities in free nerve endings without any relationship to vessels (Figure 2B). The δ-opioid receptor appeared to be more abundant in the proximal portion than in the distal portion of the paratenon.
Co-localization of Enkephalins and Their Receptors
Double staining disclosed co-existence of each of the enkephalins with DOR in the nerve fibers (Figure 3). Thus, co-localization was found in the paratenon, loose connective tissue, and the musculotendinous junction. The co-existence was found both in free nerve endings mostly localized to the paratenon and the musculotendinous junction and in vascular fibers, predominantly in the surrounding loose connective tissue. The major portions of the enkephalins and DOR identified in the nerve fibers were found to be co-localized.
Characterization of Opioid Receptors
The tissue possessed opioid binding sites, shown with the non-selective opioid ligand [3H]-naloxone. Binding of [3H]-naloxone to homogenates of achilles tendon from rat gradually increased with time and reached a steady state after 45 min at 25C (Figure 4). The binding of [3H]-naloxone was specific and saturable (Figure 5). The equilibrium dissociation constant (Kd) for the identified binding sites disclosed high affinity (Kd 7.01 ± 0.98 nM) but low binding capacity (Bmax 23.52 ± 2.23 fmol/mg protein). The Hill coefficient (nH) value was 0.89 ± 0.09.
Immunofluorescence micrographs of longitudinal sections through the achilles tendon after incubation with antisera to LE (
Immunofluorescence micrographs of longitudinal sections through the achilles tendon after incubation with antisera to DOR. Arrows denote varicosities and nerve terminals. Immunoreactivity to DOR (
Discussion
Our study convincingly shows that connective tissue is supplied with a peripheral opioid system. The neuronal occurrence of different enkephalins and the existence of δ-opioid receptors in the achilles tendon suggest that there is an anti-nociceptive system in musculoskeletal tissues, which most likely counteracts the effects of the sensory nervous system. It may prove that this can be exploited in the therapeutical setting by developing new drugs that act selectively in the periphery to mitigate pain and inflammation in musculoskeletal disorders.
Among the opioid peptides analyzed in the achilles tendon, enkephalins and DYN B represent classical opioid peptide families, whereas N/OFQ has only lately been detected and characterized (Meunier et al. 1995; Reinscheid et al. 1995; Darland et al. 1998). All opioids have been shown to have anti-nociceptive effects in animals (Millan 1986; Calo et al. 2000). Although our combined findings by IHC, RIA, and receptor binding assay indicate that there is a source of opioids in the periphery, only IHC unequivocally demonstrated their occurrence in peripheral nerves.
Immunofluorescence micrograph of a longitudinal section through the achilles tendon, showing both single and double staining with LE and DOR antibodies. Arrows denote varicosities. Immunoreactivity to LE (
Influence of incubation time on specific [3H-]naloxone binding to homogenates from rat achilles tendon at 25C. Homogenates were incubated with [3H]-naloxone (1.5 nM) for the indicated periods of time.
Saturation binding of [3H]-naloxone to homogenates from rat Achilles tendon. Representative saturation curve and Scat-chard plot (inset) (
There are two principal sources of opioid peptides in the periphery. One is represented by the immune cells, which have been shown to contain and release opioids and to be implicated in mitigating inflammatory pain (Stein et al. 1993). The other source is most likely the peripheral nervous system, about which current knowledge of opioid distribution and function is sparse. Notably, a recent study showed that approximately 17% of peripheral cutaneous axons contain enkephalins presumed to be of importance in regulating pain and inflammation (Carlton and Coggeshall 1997). The failure to detect N/OFQ in immune cells and nerve fibers by IHC, and the fact that measurable concentrations were obtained by RIA, should be attributed to the fact that RIA is a more sensitive assay. Although the median concentration of N/OFQ was below the detection limit, we still believe, on the basis of the RIA findings, that this mediator exists in connective tissue. As for DYN B, however, its nonexistence or presence in low amounts in tendon tissue under normal physiological conditions remains to be established. Notably, it has been shown that peripheral sensory nerves immunostain for dynorphin in inflammation but not under normal conditions (Hassan et al. 1992). The observed predominance of LE-positive nerves compared to the other enkephalins tested may be explained by the fact that LE is the only enkephalin processed from both proenkephalin and prodynorphin (Lewis et al. 1980; Zamir et al. 1984; Dores et al. 1990).
The physiological role of enkephalins can be assumed to be anti-nociceptive (Carlton and Coggeshall 1997; Machelska and Stein 2000), anti-inflammatory (Lembeck et al. 1982; Hong and Abbott 1995), vasodilatory (Florez and Mediavilla 1977; Moore and Dowling 1982), immunosuppressive (Brown and Van Epps 1985), and trophic (Willson et al. 1976; Zagon and McLaughlin 1991). Our observations showing that enkephalins are present in tendons and can be extracted and quantified for comparison of normal and pathological conditions comply with a recent report on RIA of bone and joint tissues (Elhassan et al. 1998). Presumably this pertains to all connective tissues. Whether the relative tissue concentrations of different opioids are altered in a specific pattern in different conditions of the locomotor system has yet to be investigated.
According to IHC the four enkephalins analyzed exhibited a similar distribution. They occurred in the paratenon and the surrounding loose connective tissue rather than in the tendinous tissue proper. Hypothetically, this difference in anatomic distribution might reflect that regulation of pain and inflammation in disorders of the achilles tendon mainly occurs in the surrounding tissues. We have previously demonstrated a similar anatomic distribution of sensory and autonomic neuropeptides in the achilles tendon, i.e., predominantly in the surrounding tissues (Ackermann et al. 1999,2001). Notably, surgical intervention in painful disorders of the achilles tendon is traditionally targeted to the tendon tissue proper without recognizing that the procedure per se entails a possible therapeutic effect by the mere cutting and denervation of the surrounding tissues.
The IHC analysis also showed that the enkephalin-positive nerve fibers were either vessel related or nonvascular. Nerve fibers around vessels can be assumed to be involved in both vasoactivity (Moore and Dowling 1982) and anti-inflammatory responses (Lembeck et al. 1982; Hong and Abbott 1995). The free nerve endings suggest a paracrine or autocrine function in the regulation of nociception. Such a regulation is probably executed in interaction with the sensory nervous system. Thus, it has been shown that the release of the sensory neuropeptide substance P (SP) from afferents in the cat knee joint is inhibited by intra-articular enkephalin-analogue injections (Yaksh 1988). Whether there is an inherent critical balance between opioids and sensory neuropeptides under normal conditions is unknown.
In this study, the existence of an opioid system in connective tissues, as demonstrated by neuronal immunoreactivity to enkephalins, is strongly supported by our opioid receptor analyses based on IHC and binding assays. Thus, δ-opioid receptors (DOR) were identified on peripheral nerve fibers in the paratenon and the surrounding loose connective tissue. Double staining for DOR and the four enkephalins tested disclosed co-existence, which complies with other studies suggesting that enkephalins are the main ligands for DOR (Dhawan et al. 1996). Our finding of enkephalins and DOR in small peripheral nerve terminals suggests localization in C-fibers, which is in agreement with other reports demonstrating that both enkephalin (Carlton and Coggeshall 1997) and DOR (Coggeshall et al. 1997; Wenk and Honda 1999) are localized in unmyelinated afferent sensory axons in the skin. However, an autonomic localization in the vessels cannot be excluded. Interestingly, DOR on peripheral nerve terminals has been shown also to co-exist with sensory neuropeptides (Wenk and Honda 1999) and to inhibit SP release (Yaksh 1988). It may prove that enkephalins acting on DOR inhibit the nociceptive and pro-inflammatory actions of sensory neuropeptides. In theory, this could involve two inhibitory mechanisms. One is inhibition of the afferent firing rate and the other is inhibition of the efferent release of sensory neuropeptides. As for the latter mechanism, DOR activity has been shown to have a very potent effect on SP release (Hirota et al. 1985; Yaksh 1988). There are several reports demonstrating that treatment with δ-opioid agonists in the periphery is both anti-inflammatory and anti-nociceptive in complex models of inflammation (Taiwo and Levine 1991; Nozaki-Taguchi and Yamamoto 1998; Zhou et al. 1998) compared to simple inflammatory models showing no anti-nociceptive effects (Levine and Taiwo 1989; Hong and Abbott 1995). The anti-nociceptive effects are mediated through DOR localized on unmyelinated afferent sensory axons (Zhou et al. 1998). Presumably, the co-existence of enkephalins and DOR in vascular nerve fibers reflects inhibition of neurogenic pro-inflammatory actions, whereas that in free nerve endings reflects inhibition of nociception.
In addition to the demonstration of δ-opioid receptors, we were also able to characterize opioid receptors in the tendon tissue by binding assays using the non-selective opioid radioligand [3H]-naloxone. The binding of [3H]-naloxone to homogenates from achilles tendon was specific, saturable, and stereo-selective. The Hill coefficient value was close to unity, indicating the non-cooperative nature of the binding process and also that [3H]-naloxone could interact with multiple classes with equal affinity.
From the combined results of our study, it seems most probable that the musculoskeletal apparatus is equipped with an opioid system. Whether there is a critical balance between the peripheral expression of opioids and sensory neuropeptides under normal circumstances, which is altered under pathologic conditions, remains to be clarified. Nonetheless, our findings appear to imply that there is a peripheral mechanism for inhibition of pain in addition to that at higher level according to the “central gate theory” postulated by Melzack and Wall (1965). It may prove that this can be exploited in the therapeutic setting by developing new drugs targeted selectively to opioid receptors in the periphery that can be administered even locally.
Footnotes
Acknowledgments
Supported by grants from the Swedish Medical Research Council (12X-08652-09B and 14X-12588-03A). Mariana Spetea was supported by fellowships from the Swedish Institute and the Wenner-Gren Foundation.