Experimental design and reporting standards for improving the internal validity of pre-clinical studies in the field of pain: Consensus of the IMI-Europain consortium.

Background and aims Pain is a subjective experience, and as such, pre-clinical models of human pain are highly simplified representations of clinical features. These models are nevertheless critical for the delivery of novel analgesics for human pain, providing pharmacodynamic measurements of activity and, where possible, on-target confirmation of that activity. It has, however, been suggested that at least 50% of all pre-clinical data, independent of discipline, cannot be replicated. Additionally, the paucity of "negative" data in the public domain indicates a publication bias, and significantly impacts the interpretation of failed attempts to replicate published findings. Evidence suggests that systematic biases in experimental design and conduct and insufficiencies in reporting play significant roles in poor reproducibility across pre-clinical studies. It then follows that recommendations on how to improve these factors are warranted. Methods Members of Europain, a pain research consortium funded by the European Innovative Medicines Initiative (IMI), developed internal recommendations on how to improve the reliability of pre-clinical studies between laboratories. This guidance is focused on two aspects: experimental design and conduct, and study reporting. Results Minimum requirements for experimental design and conduct were agreed upon across the dimensions of animal characteristics, sample size calculations, inclusion and exclusion criteria, random allocation to groups, allocation concealment, and blinded assessment of outcome. Building upon the Animals in Research: Reportingin vivo Experiments (ARRIVE) guidelines, reporting standards were developed for pre-clinical studies of pain. These include specific recommendations for reporting on ethical issues, experimental design and conduct, and data analysis and interpretation. Key principles such as sample size calculation, a priori definition of a primary efficacy measure, randomization, allocation concealments, and blinding are discussed. In addition, considerations of how stress and normal rodent physiology impact outcome of analgesic drug studies are considered. Flow diagrams are standard requirements in all clinical trials, and flow diagrams for preclinical trials, which describe number of animals included/excluded, and reasons for exclusion are proposed. Creation of a trial registry for pre-clinical studies focused on drug development in order to estimate possible publication bias is discussed. Conclusions More systematic research is needed to analyze how inadequate internal validity and/or experimental bias may impact reproducibility across pre-clinical pain studies. Addressing the potential threats to internal validity and the sources of experimental biases, as well as increasing the transparency in reporting, are likely to improve preclinical research broadly by ensuring relevant progress is made in advancing the knowledge of chronic pain pathophysiology and identifying novel analgesics. Implications We are now disseminating these Europain processes for discussion in the wider pain research community. Any benefit from these guidelines will be dependent on acceptance and disciplined implementation across pre-clinical laboratories, funding agencies and journal editors, but it is anticipated that these guidelines will be a first step towards improving scientific rigor across the field of pre-clinical pain research.

Generation and phenotypic characterization of a galanin overexpressing mouse.

In most parts of the peripheral nervous system galanin is expressed at very low levels. To further understand the functional role of galanin, a mouse overexpressing galanin under the platelet-derived growth factor-B was generated, and high levels of galanin expression were observed in several peripheral tissues and spinal cord. Thus, a large proportion of neurons in autonomic and sensory ganglia were galanin-positive, as were most spinal motor neurons. Strong galanin-like immunoreactivity was also seen in nerve terminals in the corresponding target tissues, including skin, blood vessels, sweat and salivary glands, motor end-plates and the gray matter of the spinal cord. In transgenic superior cervical ganglia around half of all neuron profiles expressed galanin mRNA but axotomy did not cause a further increase, even if mRNA levels were increased in individual neurons. In transgenic dorsal root ganglia galanin mRNA was detected in around two thirds of all neuron profiles, including large ones, and after axotomy the percentage of galanin neuron profiles was similar in overexpressing and wild type mice. Axotomy reduced the total number of DRG neurons less in overexpressing than in wild type mice, indicating a modest rescue effect. Aging by itself increased galanin expression in the superior cervical ganglion in wild type and transgenic mice, and in the latter also in preganglionic cholinergic neurons projecting to the superior cervical ganglion. Galanin overexpressing mice showed an attenuated plasma extravasation, an increased pain response in the formalin test, and changes in muscle physiology, but did not differ from wild type mice in sudomotor function. These findings suggest that overexpressed galanin in some tissues of these mice can be released and via a receptor-mediated action influence pathophysiological processes.

Antinociceptive effects after intrathecal administration of phosphodiester-, 2'-O-allyl-, and C-5-propyne-modified antisense oligodeoxynucleotides targeting the NMDAR1 subunit in mouse.

In the present study, we have compared the antinociceptive effect of three different types of antisense oligodeoxynucleotides targeting the N-methyl-D-aspartate (NMDA) R1-subunit in mice. The probes were administrated intrathecally three times during a period of 5 days (1, 5 or 25 microg/injection), followed by evaluation using the formalin test. The antinociceptive effect was correlated to in vitro receptor binding in spinal cord sections. The tissue distribution was studied after a single injection of fluorescein-conjugated probes. The phosphodiester probe showed superficial tissue penetration after 30 min and disappeared within 2 h. The probe did, however, significantly reduce both receptor binding in laminae I and II (by 36-44% compared to saline) as well as pain behavior (32% compared to saline), without apparent side effects. The mismatched probe was ineffective at 25 microg, while some reductions in receptor binding and pain behavior were seen after 5 microg. The C-5-propyne-modified phosphorothioate probe showed pronounced tissue penetration and cellular uptake as soon as 30 min after injection which was still detectable after 24 h. Immediately after injection of the highest dose, long-lasting hind-limb paralysis was observed. Receptor binding was reduced but not in a dose-related manner. Pain behavior was significantly reduced by 40% following 25 microg of antisense probe but not after lower doses or 25 microg of mismatched probe. The 2'-O-allyl-modified probe did not significantly reduce receptor binding or pain behavior. Thus, only the phosphodiester probe showed a significant correlation between reduction in pain behavior and receptor binding. These findings demonstrate that antisense technology is associated with specificity problems, but still could provide a valuable tool to study the role of different target proteins in the drug discovery process.

Assessing mechanical allodynia in the rat paw with a new electronic algometer.

Von Frey filaments used for testing mechanical thresholds are mechanically unstable and their use is difficult to standardize. We have therefore constructed a hand-held electronic pressure algometer. The pressure algometer is connected to a computerized data collection system, allowing on-line display of the applied force as well as the application rate. Data stored on the computer can be replayed and further analyzed. Using this apparatus, we have measured the pressure-induced withdrawal thresholds in rats with surgically induced neuropathy. The probe, with a circular tip of 1.0 mm diameter, was applied manually with a pressure increasing by approximately 0.05 N/s. Presurgical thresholds were normally distributed with a mean of 0.415 N, showing no significant difference between paws. During 2 weeks after surgery, the thresholds of the operated side were significantly reduced (range, 0.209-0.318 N), while the thresholds of the non-operated side remained at higher values (range, 0.432-0.491 N). Thresholds of control rats without surgery were in the 0.380-0.520 N range, with no significant difference between paws. In an additional experiment it was shown that interobserver reliability was high, both between withdrawal threshold values obtained and between rates of application used. In conclusion, the electronic algometer allows standardization of testing, detailed documentation of each experiment and provides an objective and accurate method for measuring the reactions of test animals to mechanical stimuli.

Effect of peripheral axotomy on dorsal root ganglion neuron phenotype and autonomy behaviour in neuropeptide Y-deficient mice.

The lumbar 5 (L5) dorsal root ganglia (DRGs) were studied in neuropeptide tyrosine (NPY)-deficient (-/-) and wild type (+/+) mice after unilateral sciatic nerve transection using in situ hybridization and immunohistochemistry. NPY, galanin and two NPY receptors (Y-Rs) were analyzed as well as self-mutilation behaviour (autotomy) and nociceptive thresholds. No difference between wild type and NPY-deficient mice was seen in the tail-flick or hot plate test. However, -/- mice showed a much stronger autotomy behaviour than wild type mice. NPY was not found in L5 DRGs in -/- mice, not even after axotomy. Galanin was upregulated to the same extent after axotomy in NPY-deficient and wild type mice. Y1- and Y2-R mRNAs were found mainly in small DRG neuron profiles. Both receptor mRNAs were downregulated after axotomy, to about the same extent in NPY-deficient as in wild type mice. In control and contralateral ganglia the mRNA levels of both receptors were lower in NPY-deficient mice than in wild type mice. The contralateral Y2-R mRNA levels did not reach control values in the NPY-deficient mice, as they did in the wild type mice. In both strains the Y1-R protein was decorating the somatic plasmalemma. The present results suggest that lack of NPY may cause exaggerated autotomy, a self-mutilation behaviour possibly related to pain sensation, in agreement with the described analgesic effect of NPY. Although significant differences in levels of Y1- and especially Y2-R mRNAs were observed between wild type and NPY-deficient mice, they were only moderate. These findings suggest that expression, regulation, localization and possible function of Y1- and Y2-Rs are not dependent on presence of the endogenous ligand. Also, deletion of NPY does not seem to influence the expression of the partly coexisting peptide galanin.

MK-801 neurotoxicity in the guinea pig cerebral cortex: susceptibility and regional differences compared with the rat.

N-methyl-D-aspartate (NMDA) receptor antagonists induce transient vacuole formation in neurons of the retrosplenial cortex and, after higher doses, necrosis in the same region. To our knowledge, all studies demonstrating these effects have been carried out in rats or mice. The present study investigated whether vacuolization occurs in the guinea pig, rats being used as controls. Female Dunkin-Hartley guinea pigs (age 15-18 weeks) were given a single subcutaneous injection of saline or the non-competitive NMDA antagonist dizocilpine maleate [(+)-MK-801; 1, 4, or 12 mg/kg]. Female Sprague-Dawley rats (age 16 weeks) received saline or MK-801 (1 mg/kg). Whatever the dose of MK-801, guinea pigs showed only occasional vacuolated neurons in the retrosplenial cortex. However, affected neurons (mainly large pyramidal cells of layer V) were found in the frontoparietal neocortex. The reaction was limited after 1 mg/kg, and seemed to reach a maximum at 4 mg/kg. Rats injected with 1 mg/kg MK-801 showed an intense vacuole reaction in neurons from layers III-IV of the retrosplenial cortex, but no affected neurons were noted in neocortical areas. We conclude that there are significant species differences in susceptibility to, and location of, vacuolization induced by NMDA receptor antagonists.

Two non-racemic preparations of a piperidine-based NMDA antagonist with analgesic activity.

(2R,3S,1'R)-3-(1-Hydroxy-2-phosphonoethyl)-2-piperidinecarboxyl ic acid 1 has been synthesized by two different methods. The NMDA receptor binding affinities (Ki values) are 74 nM for compound 1, and 64 nM for the corresponding ketone 2. The analgesic effects were evaluated using the mouse hot-plate test and the mouse formalin model. The ED50 values for the racemates of compounds 1 and 2, using the mouse hotplate and intrathecal injection, were 0.53 and 0.51 nmol, respectively.

(+/-)-CP-96,345, an NK1 receptor antagonist, has local anaesthetic-like effects in a mammalian sciatic nerve preparation.

The effect of the non-peptide NK1 receptor antagonist (+/-)-CP-96,345 was investigated on the compound action potential (cAP) recorded in the isolated guinea-pig nerve, and compared to the effects of the local anaesthetic lidocaine and the L-type Ca2+ channel antagonist diltiazem. (+/-)-CP-96,345, as well as lidocaine and diltiazem, produced a concentration dependent reduction of the cAP amplitude. All three drugs showed frequency dependent block. The block of the cAP by lidocaine was fully reversible at all concentrations tested, while the block by (+/-)-CP-96,345 and diltiazem were only partly reversible at higher concentrations. The present findings indicate that (+/-)-CP-96,345 exerts a local anaesthetic-like effect on nerve conduction.

Weight bearing as an objective measure of arthritic pain in the rat.

This study compares two methods for evaluating pain-related behavior in an animal model with carrageenan-induced monoarthritis. Rats injected with lambda-carrageenan into the right tibio-tarsal joint were videofilmed at various times after injection and later scored regarding their stance. Immediately after each videorecording session the animals were tested in a box constructed to register the weight load exerted by the hindpaws by means of force plates inserted in the floor. Following carrageenan injection (300 micrograms in 50 microL) the load on the injected paw fell from a control value of 39.3% +/- 0.4% of the body weight (mean +/- SEM, n = 6) to a minimum of 5.1% +/- 1.8% at 6 hr and then slowly increased to approach control levels at 72 hr. The weight load on the contralateral paw increased from a control value of 38.9% +/- 0.6% to 52.4% +/- 1.4% at 6 hr, whereafter it gradually decreased. The video-based stance scores also showed a maximal impairment at 4-6 hr, with a gradual return towards control values at 72 hr. However, the results based on the force plate measurements were less variable and more graded. Morphine inhibited the carrageenan-induced effect in a dose-dependent manner in both paradigms. In conclusion, the present results indicate that measurement of weight bearing as described in the present paper is a practical, useful, and objective method to assess the degree of arthritic pain in the rat.

Systemic or intracerebroventricular injection of NMDA receptor antagonists attenuates the antinociceptive activity of intrathecally administered NMDA receptor antagonists.

We have previously reported that the response latency in the mouse hot-plate test is affected differently by spinal intrathecal (i.t.) injection of competitive and non-competitive N-methyl-D-aspartate (NMDA) receptor antagonists, in that only the former produces an antinociceptive effect. Since the lipophilic non-competitive antagonists will redistribute rapidly from the spinal injection site, it is conceivable that they reach sites where they counteract the spinal antinociceptive effect. In the present study, we have tested this hypothesis by comparing the antinociceptive effect of the competitive NMDA receptor antagonist CGS 19755 and the non-competitive NMDA receptor antagonist MK-801 after i.t., intraperitoneal (i.p.) and intracerebroventricular (i.c.v.) administration as well as after combinations thereof. CGS 19755 injected i.p. or i.c.v. and MK-801 injected i.p. or i.t. attenuated the antinociceptive effect of i.t. injected CGS 19755. Both i.p. and i.c.v. administration of either CGS 19755 or MK-801 dose-dependently impaired motor function without producing antinociceptive effects. Thus, the effect of CGS 19755 and MK-801 on the motor system was found to be separate from their antinociceptive effect. In a separate experiment, changes in hind-paw skin temperature were excluded as a possible confounding factor. These findings demonstrate that supraspinal systems can limit the spinal antinociceptive effect of NMDA receptor antagonists.

Supraspinal mediation of dopamine-serotonin interactions in extrapyramidal motor functions in the rat.

The effects of intracerebroventricularly (i.c.v.) or intrathecally (i.t.) administrated 8-OH-DPAT on catalepsy, induced by the specific DA D2 antagonist raclopride (16 mg kg-1 s.c.), were studied in rats. It was found that 8-OH-DPAT (0.5 or 2.0 micrograms kg-1) injected by the i.c.v. route produced a statistically significant of raclopride-induced catalepsy at both doses. In contrast, 8-OH-DPAT (0.2 or 2.0 micrograms kg-1) given by the i.t. route had no statistically significant effect on the raclopride-induced catalepsy. These results suggest that the antagonistic effect of 8-OH-DPAT on catalepsy, induced by DA blocking agents, is primarily mediated at the supraspinal level.

Tissue distribution of two NMDA receptor antagonists, [3H]CGS 19755 and [3H]MK-801, after intrathecal injection in mice.

The tissue distribution of [3H]cis-4-phosphonomethyl-2-piperidine carboxylic acid (CGS 19755) and [3H](+)-5-methyl-10,11-dihydro-5H-dibenzo[a,d]cyclohepten-5,10-imi ne (NK-801) was investigated after a single IT injection into lumbar spinal cord of mice. The level of radioactivity was analyzed in the lumbar, thoracic, and cervical spinal cord, brainstem, frontal cortex, liver, lungs, kidneys, stomach, intestine, spleen, heart, and blood from 5 min up to 6 h after injection. Within the CNS, [3H]CGS 19755 redistributed slowly from the site of injection toward the brainstem and cortex, peaking in the cortex 3-4 h after IT injection. At no time, however, did the relative level per gram of tissue in the frontal cortex exceed 10% of the relative level in the lumbar region of the spinal cord. The highest peripheral level of [3H]CGS 19755 was found in the kidneys. [3H]MK-801 redistributed rapidly from the spinal cord injection site to the peripheral organs. The highest peripheral levels of [3H]MK-801 were found in the lungs and liver, where the radioactivity peaked at 10 and 30-60 min, respectively, after injection. The relative levels of [3H]CGS 19755 were consistently higher in CNS tissues (except for the first 15 min in the frontal cortex) and blood than the corresponding levels of [3H]MK-801. The opposite relationship was true in the liver, lungs, kidneys, stomach, intestine, spleen, and heart. The effect on the response latency in the hot-plate test was quantified in the same animals immediately prior to sacrifice for the distribution study.(ABSTRACT TRUNCATED AT 250 WORDS)

Potentiation of a behavioural response in mice by spinal coadministration of substance P and excitatory amino acid agonists.

The functional interaction in the spinal cord between substance P and excitatory amino acid agonists was investigated. Behavioural responses were scored in mice after intrathecal administration of excitatory amino acid agonists and substance P, given separately or in combination. A strong potentiation of the effect was seen when substance P was coadministered with N-methyl-D-aspartate (NMDA), alpha-amino-3-hydroxy-5-methyl-4-isoxazole-propionic acid (AMPA) or kainic acid (KA). The potentiation was blocked by the corresponding antagonists: the selective NMDA-receptor antagonist (+/-)-3- (2-carboxypiperazin-4-yl)propyl-1-phosphonic acid (CPP), the non-NMDA receptor antagonist 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX) and the substance P analog, [D-Arg1,D-Trp7,9,Leu11]-substance P (Spantide). These findings indicate a functional interaction between substance P and glutamate in the dorsal horn of the spinal cord, compatible with the hypothesis that corelease of substance P and glutamate from primary afferent neurons may enhance nociception.

Hyperbaric exposure and morphine alter the pattern of behavior in the formalin test.

This study investigates the behavioral effects of morphine administration and exposure to high ambient pressure in the formalin test. Rats were simultaneously given formalin (0.1 ml, 5%) in a hind paw, and saline or morphine (2.5-10.0 mg/kg) subcutaneously. They were then exposed to ambient pressure of either 1 or 48 bar (compression rate: 3 bar/min; 1 bar is approximately equal to the pressure of 1 atmosphere) in a helium-oxygen atmosphere. The behavior of the animals was monitored for 35 min at stable pressure, starting 25 min after the injections. After morphine, the groups tested at 1 bar showed a dose-dependent reduction in pain-related activities such as licking, biting, clutching and protecting the injected paw but paw-elevation while resting was significantly increased after the highest dose. The 48-bar groups spent almost no time in these behavioral categories but showed an increase in apparently normal motor activity. Paw-jerking appeared to be a more robust response. The number of jerks was not altered by pressurization and was dose-dependently reduced by morphine at both pressures. The results show that hyperbaric exposure alters the response pattern in the formalin test, demonstrate the advantage of evaluating several behavioral criteria in this test and provide tentative evidence against pressure reversal of morphine analgesia.

Lesions of bulbo-spinal serotonergic or noradrenergic pathways reduce nociception as measured by the formalin test.

Intrathecal administration of the neurotoxins 5,6-dihydroxytryptamine (5,6-DHT) and 6-hydroxydopamine (6-OHDA) in rats selectively lesioned the descending spinal serotonergic and noradrenergic pathways, respectively. Four days after neurotoxin administration the behaviour was evaluated in the formalin test. Several behavioural variables were recorded. The statistical analysis of the results was supplemented using a multivariate statistical method (partial least squares projection to latent variables, PLS) in addition to traditional analysis of variance. The described methods for recording and statistical analysis of behaviour appear to be useful in describing drug-induced differences in behavioural patterns in the formalin test. Both types of lesion reduced the pain-related behaviour in the formalin test (protection of the paw, biting and licking). The results indicate that the descending monoaminergic pathways are parts of a network which maintains adequate nociceptive responses to a chemical stimulus, or to stimuli lasting several minutes, as in the formalin test.

Effects of ethanol on body temperature of rats at high ambient pressure.

Separately, ethanol and high ambient pressure cause hypothermia in laboratory animals. However, ethanol and high pressure have mutually antagonistic effects on several biological functions and the present experiments investigate their combined action on body temperature. Rats given saline, 1.5 g/kg ethanol or 3.5 g/kg ethanol were exposed to 1 bar air at 25-26 degrees C, 1 bar helium-oxygen at 30-31 degrees C, or 48 bar helium-oxygen at 33.5-34.5 degrees C. Ambient, colonic and tail-skin temperatures were monitored for 60 min. There were no significant differences in mean ambient or tail-skin temperatures between groups belonging to the same ambient condition. Colonic temperatures under the 1 bar conditions were 1.5-2 degrees C lower in the 3.5 g/kg ethanol group than in the saline and 1.5 g/kg ethanol groups, while no significant differences were observed between the groups at 48 bar. Comparisons of the colonic temperatures at the end of the observation period, i.e., 60 min after administration of ethanol, demonstrated that their values at 48 bar were significantly lower than at 1 bar after saline, significantly higher after 3.5 g/kg ethanol and identical across conditions in the 1.5 g/kg groups. The results suggest that high ambient pressure may counteract rather than potentiate the hypothermic effect of ethanol.

The increasing-temperature hot-plate test: an improved test of nociception in mice and rats.

The increasing-temperature hot-plate test has several advantages compared to the conventional hot-plate test, but available equipment has been impractical and restricted with regard to stimulus control. We now describe an apparatus consisting of an aluminum plate that is heated and cooled by Peltier elements in contact with its lower surface. Several plates can be used simultaneously, individually controlled by electronic proportional feedback circuits. The set temperature of the feedback circuit is controlled by a computer program run on an IBM XT-compatible PC, so that a linear increase in temperature is achieved. Experiments were performed using rats and mice, with hindpaw licking as an end-point criterion. Experiments with various heating rates showed that 3.0 degrees C/min is the lowest rate that can be applied without signs of stress in the animals. On the basis of the recorded data, nociceptive temperature thresholds were calculated to be approximately 44.5 degrees C for both rats and mice. Inspection of the paws after analgesic treatment and exposure to different end-point temperatures suggested that a cutoff temperature of 50 degrees C should be employed to minimize tissue damage. Testing at ambient temperatures of 18 degrees and 28 degrees C yielded similar results for rats, whereas mice responded at significantly higher plate temperatures in the colder environment. Dose-related antinociceptive effects were demonstrated for morphine and paracetamol in both species. The results confirm that the increasing-temperature hot-plate test is a valuable test of nociception, which is also suitable for demonstrating the antinociceptive effects of nonopioid analgesics. The test may also be used to estimate the nociceptive temperature threshold.

Interaction of high pressure and a narcotic dose of ethanol on spontaneous behavior in rats.

This study analyses the spontaneous motor activity of rats that had received a narcotic dose of ethanol (3.5 g/kg) and were then exposed to 1 atmosphere absolute pressure (ATA) air or to 1 or 72 ATA of helium-oxygen (heliox). The ambient temperature was adjusted to offset ethanol-and helium-induced hypothermia. Ethanol administration prevented the occurrence of convulsions but did not alter the total number of myoclonic jerks at stable pressure. The ethanol-intoxicated animals exposed to high pressure did not exhibit normal locomotion but showed a trend towards increased activity during the last observation period. Similar blood and brain concentrations of ethanol were found in the 1 and 72 ATA groups. These results show that exposure to 72 ATA for 40 min started to exert some antagonistic effects, and they suggest that exposure to higher pressures or for a longer period of time may be sufficient to significantly offset the depressant effects of a narcotic dose of ethanol on spontaneous behavior in rats. At the same time, ethanol seems to protect against some aversive effects of high pressure.

Interaction of ethanol and the high pressure nervous syndrome in rats.

The aim of this study was to determine whether administration of ethanol protects rats against the preconvulsive symptoms of high pressure nervous syndrome (HPNS). Male Sprague-Dawley rats were given either saline or 0.5, 1.5 or 2.5 g/kg ethanol i.p. After injection, the animals were individually exposed to helium-oxygen at 60 atmospheres absolute (atm abs) pressure. The chamber temperature was adjusted to counteract ethanol- and helium-induced hypothermia. Several behavioral parameters were scored continuously during the first 64 min after injection. The time spent in tremor at high pressure was significantly less in the 1.5 and 2.5 g/kg ethanol-treated groups. The number of jerks was significantly lower in the 2.5 g/kg ethanol-treated group. The two highest doses of ethanol induced a characteristic pattern of unsteady locomotion, which was returned to normal in the 1.5 g/kg group at 60 atm abs. Other behavioral effects of ethanol, such as depression of total motor activity, were also reduced. These results indicate that ethanol can significantly ameliorate some of the adverse symptoms of HPNS in freely moving rats.