Non-opioid peptides for analgesia J. Chrubasik, S. chrubasikl and E. Martin Department of Anesthesiology, University Hospital, Im Neuenheimer Feld, D-69 Heidelberg and '~e~artmentof Forensic Medicine, University of Freiburg, Albertstrasse, D-78 Freiburg, Germany Abstract. Amongst the spinal peptide candidates believed to be involved in the mediation of analgesia, only somatostatin fulfills the criterium of a real analgesia substance. Spinal somatostatin specifically blocks the transmission of painful stimuli. Spinal calcitonin may lower the opioid dose requirement in patients with bone metastases but it fails to relieve acute pain. The usefulness of ACTH and CRF for treatment of pain remains to be established. The role of CCK-8, vasopressin and neurotensin is unclear. The contradictory findings on antinociception using simple rodent withdraw1 reflex tests (e.g. the tail flick test), or more complex behavioral tests in which supraspinal sensory processing is involved, (e.g. the hot plate test), indicate that these tests are inappropriate when neuropeptides are employed. Furthermore, due to their inability to predict analgesia in humans, they do not fulfill the guidelines proposed by the IASP that animal test procedures have to be Address for correspondence: for the benefit of humans. Joachim Chrubasik Department of Anesthesiology University of Heidelberg Im Neuenheimer Feld 110 Key words: ACTH, CRF, CCK-8, vasopressin, neurotensin, calcitonin, D-69 Heidelberg, Germany somatostatin, spinal analgesia 290 J. Chrubasik et al. INTRODUCTION CHOLECYSTOKININ OCTAPEPTIDE (CCK-8) Animal studies indicate that non-opioid peptides are involved in the modulation of sensory processes Cholecystokinin immunoreactivity is widely including pain transmission. Cholecystokinin octa- distributed within the central nervous system, peptide, neurotensin, vasopressin, adrenocortico- e.g. in the cerebral cortex, striatum, hippocampus, tropic hormone (ACTH), corticotropin-releasing amygdala and parts of the brain stem. Receptor factor (CRF), calcitonin and somatostatin are the autoradiographic studies suggest that receptors for most promising non-opioid candidates that produce CCK-8 are located on axons that may act pre-synap- analgesia. tically to modify the input of sensory information Unfortunately, only few clinical studies give (Innis and Aghajanian 1984). Table I summarizes ample evidence for the non-opioid analgesic effect the effect of CCK-8 on nociceptive thresholds in a in humans. Contradictory results from rodent test variety of common rodent test procedures. Con- procedures make it difficult to decide whether the trasting results were obtained in the rodent tests des- one or other non-opioid is an analgesic in man or not. pite same CCK-8 doses and routes of administration TABLE I Effects of CCK-8 in common rodent test procedures, the dosages employed, routes of administration (sc, subcutaneously; ith, intrathecally; icv, intraventricularly), antinociceptive findings and references Antinociception Rodent Tests Yes Reference No Reference Mouse Hot Plate Jump 50- 1,000 pgkg sc Zetler 1980 50-500 pgkg sc Hill et al. 1987 300 pgkg sc Barbaz et al. 1985 2 pg icv Lick 3-30 pg icv Hill et al. 1987 0.003-0.3 pg icv Hill et al. 1987 10,000 pgkg sc Mouse Tail Flick 50-750 pgkg sc Zetler 1980 1 pg icv Kubota et al. 1987 11-34 ng icv Barbaz 1986 Rat Hot Plate Jump 0.8-3.2 pg icv Itoh et al. 1982 Rat Paw Pressure 10-225 ng ith Pittaway et al. 1987 2250-15000 ng ith Pittaway et al. 1987 3300 pgkg sc Hill et al. 1987 40-3000 pgkg sc Hill et al. 1987 Mouse Writhing Acet Acid 15-60 ng ith Hong and 120- 1920 ng itch Hong and 15-60 icv Takemori 1989 Takemori 1989 750 pgkg sc Zetler 1980 Phenylchinon 4.5 pgkg icv Barbaz et al. 1986 3 ~gkgsc 2,000 pgkg sc Acetylcholin 1,100-10,000 pgkg sc Hill et a1. 1987 Non-opioid peptides for analgesia 291 and even when the same test procedure was used, sensory functions (Faull et al. 1989). A naloxon-ir- e.g. the mouse hot plate test (Hill et al. 1987, Zetler reversible analgesia was achieved in the rat hot plate 1980) or the rat paw pressure test (Hill et al. 1987). and tail flick tests following administration of neur- Baber et al. (1989) postulated that "pharmacologi- otensin into the periaqueductal grey. Electrophysi- cal" doses of CCK-8 in contrast to small CCK-8 ological studies revealed that the analgesic effect doses produce analgesia. However, part of the was elicited by excitation of neurons that activate CCK-8 doses that produced antinociception in the descending inhibitory nerve fibres to dorsal horn rat tail flick, rat paw pressure and mouse writhing neurons involved in the mediation of pain (Behbehani tests were lower (Hong and Takemori 1989; Jurna and Pert 1984). Similarly, direct microinjection of and Zetler 198 1, Pittaway et al. 1987) than the doses neurotensin into the central nucleus of the amygdala that produced hyperalgesia or were without effect produced a significant increase in the antinocicep- (Table I). The fact that the results obtained in most tive threshold when using the rat hot plate test. Le- studies could not be reproduced by other groups left sions of the stria terminallis totally abolished this the question unanswered as to whether CCK-8 is an antinociceptive effect (Kalivas et al. 1982). analgesic substance or not. However, equal intra-cerebroventricular doses Rodent test procedure (the rat hot plate and tail of neurotensin produced antinociception in the flick tests) give evidence that the CCK-8 antagonist mouse hot plate test (Osbahr et al. 1981), but not in proglumide potentiates opioid analgesia, even the rat hot plate test (Kalivas et al. 1982). Further- when administered systemically (Barbaz et al. more, the antinociceptive response in the rat tail 1985). It was, however, also suggested that CCK-8 flick test with equivalent intra-cerebroventricular might antagonize opioid analgesia (Katsuura and doses of neurotensin differed (Clineschmidt et al. Itoh 1985, Watkins et al. 1985). Three clinical in- 1979, Pazos et al. 1984) (Table 11). The rodent test vestigations produced further contradictory find- procedures did, therefore, not clearly demonstrate ings on the effect of proglumide on morphine the neurotensin antinociceptive effect. analgesia in humans. Whereas 50 pg intravenous proglumide potentiated the analgesic effect of mor- phine after inducing experimental pain (Price et al. VASOPRESSIN 1985) or following surgical removal of all four third molar teeth (Lavigne et al. 1989), 0.5 pg, 100 pg The pituitary polypeptide hormone vasopressin and 50 mg intravenous proglumide have proved to is present in the brain (Sofroniew 1980) and cere- be ineffective in potentiating morphine analgesia brospinal fluid (Jenkins et al. 1980). The release of for the alleviation of pain after abdominal oper- vasopression in man is controlled by endogenous ations (Lehmann et al. 1989). It therefore seems un- opioids (Rossier et al. 1979). This may suggest a likely that the neuromodulatory role of CCK-8 in linkage between vasopressin and pain. The contra- antinociception (Wiesenfeld-Hallin and Duranti dictory results when equal doses of intrathecal va- 1987) is of clinical relevance. sopressin were employed in the rat tail flick test (Millan et al. 1980, Thurston et al. 1988) make it NEUROTENSIN difficult to predict whether intrathecal vasopressin is an analgesic in humans (Table 11). The tridecapeptide neurotensin is mainly dis- tributed throughout the periaqueductal gray, sub- ACTH stantia gelatinosa and locus coeruleus (Uhl et al. 1979). The anatomical localization of neurotensin The pituitary peptide ACTH acts like an agonist- receptors in the deeper inner segment of the spinal antagonists at central opioid receptors. Opioid in- cord also indicate their involvement in modulating duced analgesia in the rat hot plate and tail-flick 292 J. Chrubasik et al. TABLE I1 Effects of neurotensin and vasopressin in common rodent test procedures, the dosages employed, routes of administration (ith, intrathecally; icv, intra-ventricularly), antinociceptive findings and references - Antinociception Rodent Tests Yes Reference No Reference Neurotensin Mouse Hot Plate 1-30 pg icv Osbarh et al. 1981 Rat Hot Plate 2.5 pg icv Kalivas et al. 1982 Rat Tail Flick 5.4 pg icv Pazos et al. 1984 2.4 pg icv Clineschmidt et al. 1979 Vasopressin Rat Tail Flick 2.5 and 25 ng itch 0.1-20 pg itch Millan et al. 1984 tests was reversed following intracerebroventricular Intracerebroventricular administration of up to 30 pg administration of low ACTH doses (Fratta et al. CRF has failed to produce analgesia in the hot plate 1981, Smock and Fields 1981). In contrast, far (Sherman and Kalin 1986) and the tail flick (Ayesta higher doses produced analgesia (Bertolini et al. and Nikolarakis 1989) tests. 1979). The effect of intrathecal ACTH, however, However, intravenous CRF (12.5 pglkg) in- was variable in that only 50 % of the rats showed a creased the latency in the rat hot plate test (Hargraeves slight increase in the antinociceptive threshold; 50 % et al. 1987). Moreover, the analgesic effect of CRF was of the rats showed a decrease or no effect (Belcher demonstrated in 14 patients having bilaterally syrnrne- et al. 1982). Small doses of intravenous ACTH en- trical impacted third molars. At each procedure an hanced jumping in a modified hot plate test, where- upper and lower third molar on one side was extracted as high ACTH doses slightly, though not significantly under local anesthesia consisting of 2 % mepivacaine. suppressed the response. However, the small Sixty minutes following surgery, subjects received in- ACTH dose which itself produced significant travenously either 1 pglkg human CRF or placebo. At hyperalgesia, prevented the analgesic effect of a second operation approximately 2 weeks later the al- 0.625 mg/kg of morphine (Amir and Amit 1979). ternative treatment was administered. Mean pain A physiological role of ACTH in antinociception scores were significantly lower at 90 and 120 minutes was recently questioned.