Intrathecal urocortin I in the spinal cord as a murine model of stress hormone-induced musculoskeletal and tactile hyperalgesia

Stress alters pain sensitivity. For example, acute restraint of rodents causes thermal antinociception ( Imbe et al. , 2004 ) whereas chronic exposure to a daily restraint causes long-lasting thermal hyperalgesia ( Okano et al. , 1995 ; Gamaro et al. , 1998 ; da Silva Torres et al. , 2003 ). Forced swims enhance thermal nociception ( Quintero et al. , 2000 ; Imbe et al. , 2004 ; Suarez-Roca et al. , 2006a ; Suarez-Roca et al. , 2006b ), chemical nociception ( Quintero et al. , 2003 ; Suarez-Roca et al. , 2008 ; Imbe et al. , 2010 ; Quintero et al. , 2011 ) and mechanical nociception in rats ( Suarez-Roca et al. , 2006a ; Okamoto et al. , 2012 ). Corticotropin-releasing factor (CRF) and its related proteins may contribute to these effects as CRF influences nociception ( Lariviere & Melzack, 2000 ; Imbe et al. , 2006 ). Consistent with this possibility, we previously demonstrated that musculoskeletal hyperalgesia is induced in mice by a forced swim and attenuated by a CRF2 receptor antagonist ( Abdelhamid et al. , 2013 ). CRF and its congeners are distributed in pain-relevant sites in the rodent central nervous system (CNS), including the insular cortex, hypocampal formation, thalamus, hypothalamus, periaquaductal gray and locus coeruleus ( De Souza et al. , 1985 ; Chalmers et al. , 1995 ; Bittencourt et al. , 1999 ), spinal cord ( Merchenthaler et al. , 1983 ; Korosi et al. , 2007 ; Kim et al. , 2011 ) and dorsal root ganglia ( Kim et al. , 2011 ). Inhibition of swim stress-induced musculoskeletal hyperalgesia ( Abdelhamid et al. , 2013 ) and attenuation of footshock-induced bladder hypersensitivity ( Robbins & Ness, 2008 ) by intrathecally injected CRF2 receptor antagonists suggests that either urocortins or CRF are released from spinal projections in response to stress. Immunofluorescence of CRF- and urocortin I-containing fibers in the spinal cord are consistent with neuronal projections from which these hormones may be released onto neuronal cell bodies expressing both CRF1 and CRF2 receptors ( Korosi et al. , 2007 ). CRF-like ligands have distinct pharmacological profiles for each receptor. CRF binds to CRF1 receptors with high affinity but to CRF2 receptors with significantly lower affinity ( Perrin et al. , 1995 ). Urocortin I binds to CRF1 and CRF2 receptors with about equally high affinity ( Chen et al. , 1993 ; Lovenberg et al. , 1995 ) but urocortin II and III have high affinity for only CRF2 receptors ( Lewis et al. , 2001 ; Reyes et al. , 2001 ). In the CNS, CRF1 and CRF2 receptors frequently have opposite effects ( Coste et al. , 2001 ; Reul & Holsboer, 2002 ) and together shape responses to stress. CRF2 receptor expression along nociceptive pathways is upregulated after neuronal damage that leads to neuropathic pain ( Kim et al. , 2011 ). Additionally, in patients with fibromyalgia syndrome CRF is elevated in the cerebrospinal fluid where its concentration correlates with pain sensitivity ( McLean et al. , 2006 ). While the antinociceptive and hyperalgesic effects of CRF have been widely studied, the effect of urocortins on nociception is not well characterized. We determined whether any of three urocortin congeners are sufficient to recapitulate stress-induced hyperalgesia in musculoskeletal and tactile nociceptive assays. Because CRF and urocortins induce release of corticosterone and influence feeding behavior ( Zorrilla et al. , 2003 ) and body temperature ( Turek & Ryabinin, 2005 ; Telegdy et al. , 2006 ), we used these parameters to validate doses and delivery routes used.

Adult Swiss-Webster female mice weighing 20–30 g (Harlan Sprague Dawley Inc., Indianapolis, IN) were used. Animals were housed five per cage, and allowed to acclimate for at least 1 week prior to use. Female mice were used because females are generally more sensitive to pain and to the effect of stress on pain ( Fillingim et al. , 2009 ) than males. Estrogen regulates promoter activity of the urocortin gene ( Haeger et al. , 2006 ), suggesting that some effects of urocortin may be more prominent in females. In addition, CRF lowers the firing rate of nociceptive-sensitive fibers in females and pre-adult males in locus ceruleus neurons ( Borsody & Weiss, 1996 ). Mice were allowed free access to food and water, and housed in a room with a constant temperature of 23°C on a 12-h light: 12-h dark cycle; light was on from 6 am to 6 pm. Experiments started around 8 am with baseline measurements; urocortin analogs and other drugs were injected at 9 am and the last behavioral test (8 h after injection) was performed at 5 pm. Animals were used strictly in accordance with the Guidelines of International Association for the Study of Pain and the University of Minnesota Animal Care and Use Committee, and those prepared by the Committee on Care and Use of Laboratory Animals of the Institute of Laboratory Animal Resources, National Research Council (DHEW Publication (NIH) 78-23, revised 1995). All efforts were made to minimize animal suffering, to reduce the number of animals used, and to utilize alternatives to in vivo techniques, if available. CRF (>95%HPLC), urocortin I (>97% HPLC), and urocortin II (>97% HPLC) were purchased from Sigma-Aldrich (St Louis, MO). NBI-35965 (>98% HPLC), astressin 2B (>98% HPLC) and astressin (>90%HPLC) were purchased from Tocris Bioscience (Ellisville, MO). We used 1–10 μg of CRF, urocortin I, and urocortin II to test whether agonists of either CRF1 and/or CRF2 receptors could influence pain when administered peripherally or delivered intrathecally (i.t.) or intracerebroventricularly (i.c.v.) directly into the CNS. This dose range (1–10 μg) was chosen because low doses (3 μg or less) were reported to cause either no effect or analgesia in other modalities of nociception ( Lariviere & Melzack, 2000 ) whereas centrally administered CRF at a dose of 25 μg can cause seizures and/or behavioral problems ( Ehlers et al. , 1983 ). However, when we found that 10 μg of urocortin I produced a large effect on mechanical nociception, 3 and 1 μg of urocortin I were also tested to examine the dose-relationship of their responses. The doses of the antagonists used (50 μg NBI-35965, 10 μg astressin 2B and 10 μg astressin) have been previously shown to be centrally active ( Martinez et al. , 2004b ) and were dissolved in double-distilled water. Morphine sulfate (>98% HPLC), obtained from Mallinckrodt (St. Louis, MO) was used for its antinociceptive activity to differentiate the effect of urocortin I on strength, which is insensitive to opiates, from its ability to induce hyperalgesia, which is inhibited by morphine ( Kehl et al. , 2004 ). Morphine was given at doses of 3, 10 and 30 mg/kg intraperitoneally (i.p.) 30 min before the von Frey fiber assay as those doses are sufficient to inhibit the effects of other compounds, like lipopolysaccharides, that induce on musculoskeletal hyperalgesia measured using the grip force assay ( Kehl et al. , 2004 ). Drugs were delivered either subcutaneously in the scruff of the neck and upper back, intraperitoneally in the lower abdominal quadrants, intrathecally or intracerebroventricularly (i.c.v.). Injections made i.t. in mice were delivered at approximately the L5–L6 intravertebral space using a 30-gauge, 0.5 inch disposable needle on a 50 μl Luer tip Hamilton syringe in lightly restrained, unanaesthetized mice ( Hylden & Wilcox, 1981 ). A volume of 5 μL was consistently used to ensure that the drug did not reach the basal cisterns of the brain, ensuring a selective distribution to the spinal cord area (Fairbanks et al., 2003) I.c.v. injections were made freehand in mice that were anesthetized using isoflurane, a drug that allows a rapid recovery from anesthesia. Proficiency in proper placement was ensured via preliminary studies in which the free flow of fluid marks injection into ventricles rather than tissues. Injections were delivered using a 30-gauge needle into the lateral ventricle using anatomical coordinates from bregma (+0.3mm anterior, −1.0mm lateral and −3mm ventral) as described in previous studies ( DeVos & Miller, 2013 ). Drugs delivered centrally were dissolved in 5 μl as this volume was found to be large enough to be easily and accurately measured. CRF and urocortin II were dissolved in saline whereas urocortin I was dissolved in saline that was acidified using acetic acid in a final concentration of 0.25%. Control mice were injected with the same volume of vehicle. Nociception was measured as described below for 3 days prior to injections to establish the baseline response and over a 24–72 h period after injections. This time-course was implemented based on the effects induced by both CRF and urocortins on other parameters widely reported in the literature. In experiments to determine the involvement of the different CRF receptors in the urocortin I-induced mechanical hyperalgesia, we tested the effect of a CRF1 receptor antagonist (NBI-35965 at 50 μg/5 μl i.t. ( Hoare et al. , 2003 )) and of a CRF2 receptor antagonist (astressin 2B at10 μg/10 μl i.t), on urocortin I-induced hyperalgesia (3 μg/5 μl i.t.). 10 μg urocortin I was delivered i.t. in a 5-μl volume 15 min after injection of the antagonist (either NBI-35965 or astressin 2B). Astressin (10 μg/10 μl i.t. or 10 μg/200 μl s.c.), a non-selective antagonist, was injected 30 min before 3 μg or 10 μg of urocortin I. Mechanical sensitivity to a #4.17 von Frey fiber (1.4 g) was measured in mice daily for at least 3 days prior to injections and typically 4, 8 and 24 h afterwards. Mice were randomized so the average response to the von Frey fiber prior to injection did not differ amongst groups. To test, mice were placed on a wire mesh grating under a glass 6-ounce (177 ml) custard cup, which prevents escape but allows movement of all four limbs and head. The von Frey fiber was applied ten times to the plantar surface of each hind paw, to the point of bending. A positive response was defined as a brisk lifting, shaking or licking of the paw. The number of positive responses out of 20 was recorded as the mechanical sensitivity. We selected the #4.17 fiber as prior tests indicated that this size produced positive responses but of a sufficiently small magnitude that they could still be potentiated to reflect hyperalgesic responses (enhanced nociception). Fiber sizes 4.08 and greater are probably nociceptive as responses to this size in preliminary studies were inhibited by pretreatment with 10 mg/kg of morphine delivered 30 min prior to von Frey fiber testing. Larger fiber sizes were not used as they produced a sufficiently large nociceptive response that hyperalgesia would not be detected due to the ceiling effect. To assess changes in musculoskeletal hyperalgesia or weakness, forelimb grip force was measured using a grip force analyzer, as described in detail for use with mice ( Wacnik et al. , 2003 ; Kehl et al. , 2004 ; Kovacs et al. , 2008 ). Hyperalgesia was differentiated from weakness by the ability of an analgesic compound (morphine) to reverse decreases in grip force caused by enhanced nociception while weakness is unaffected. The apparatus consists of a force transducer that is connected to a wire mesh grid (12 × 7 cm 2 O.D. with a ~0.5 cm square wire grid) and is positioned on top of an aluminum frame approximately 30 cm above the bench top. During testing, each mouse was held by its tail (close to the base) and gently passed in a horizontal direction over the wire grid until it grasped the grid with its forepaws. The peak force exerted by the forelimbs of each mouse on the grid was recorded by a force transducer. Three grip force measures were obtained, and the average of these measurements was used to represent each animal’s forelimb grip force. Animals were familiarized with the grip force apparatus for two days prior to the recording of baseline values and injection of stress hormones on the third day, one hour prior to administration of drugs. The mean grip force values obtained 4, 8 and 24 h after injections are represented for each group. We monitored the efficacy of CRF and urocortin I and II on body weight and temperature to assess their effects on eating and thermoregulation. Body weight was determined using an A&D FX-300i top-loading balance from Cole-Palmer (Vernon Hills, IL). The rectal temperature was measured using a digital thermometer (model Micro Therma 2-ETI, Braintree Scientific Inc., Braintree, MA). When comparing two treatments, we used an unpaired, two-tailed Student’s t-test. Differences in body temperature and body weight at specific time-intervals after drug injection were compared to their appropriate pre-injection control values using a paired two-way Student’s t -test. Data involving comparisons of more than two groups were analyzed using one-way ANOVA followed by the Newman-Keuls post hoc analysis to determine differences amongst treatment groups tested at the same time or amongst the areas under the curves of multiple treatment groups. Statistical analyses were performed on data derived from the same experiments but figures in some cases depict pooled control values to simplify graphs. Where indicated, the mean grip force values and von Frey fiber responses shown are taken during the peak effect on mechanical sensitivity (where the value for each animal was the average of those taken 4 and 8 h after injections) to capture the optimal hyperalgesic effect of urocortin I. In all analyses, differences were considered significant if the probability that it occurred because of chance alone was less than 5% ( P <0.05). All statistics were completed using Graphpad-Prism Graphpad Software Inc., La Jolla CA, USA.

The effects produced by injection of 10 μg of urocortin I (i.t.) on both von Frey fiber and grip force responses are shown in figure 1 for comparison to it’s congeners and to other routes of administration. To simplify comparisons between treatments and routes of injection, the effects of urocortin I, urocortin II and CRF were calculated and depicted as areas under the curve from 0 to 24 h compared to that of their respective vehicle-injected control groups ( Fig. 1A–H ). This time-interval was selected for analysis as it captured the full effect of urocortin I in the grip force assay. When injected intracerebroventriculary (10 μg in 5 μL), urocortin I, urocortin II and CRF each failed to influence either von Frey (urocortin l: t 18 =1.62, P =0.12; urocortin II: t 12 =0.07, P =0.95; CRF: t 12 =0.37, P =0.71) or grip force responses (urocortin 1: t 18 =1.29, P =0.22; urocortin II: t 12 =0.19, P =0.86; CRF: t 12 =0.08, P =0.93, as the areas under the curve for von Frey fiber and grip force responses did not differ from their respective vehicle-injected controls. However, when this dose was injected intrathecally, urocortin I significantly increased responses to the von Frey fiber ( Fig. 1A ; t 39 =3.91, P =0.0003) and decreased muscular force exertion as recorded by the grip force assay ( Fig. 1 E ; t 55 =3.22, P =0.002). The area under the curve for the longer-lasting effect of 10 μg of urocortin I on von Frey fibers was also significantly greater than that of controls when analyzed over a 48-h time-interval ( Fig. 2A ; t 61 =3.74, P =0.0004) that captures the full time-course of the effect of this dose on von Frey fiber responses. Based on their relatively greater selectivity, we then tested CRF (selective for CRF1) and urocortin II (selective for CRF2) for their effects on von Frey and grip force responses when compared to urocortin I. When injected at the same high dose (10 μg) and by the same route (i.t.) as that producing a significant and long-term hyperalgesia following urocortin I ( Fig. 2 ), CRF and urocortin II each failed to replicate the hyperalgesic effect of urocortin I when measured using either the von Frey fiber ( Fig. 1B ; t 31 =0.70, P =0.49; Fig. 1C ; t 31 =1.27, P =0.21) or grip force assay ( Fig. 1F ; t 31 =1.00, P =0.32; Fig. 1G ; t 31 =0.04, P =0.97). One difference between the injection of urocortin I compared to those of urocortin II or CRF is that the latter two were not injected in an acidified vehicle as they did not require acidification to dissolve. To determine whether the failure of CRF and urocortin II to cause hyperalgesia was due to a lack of co-activation of acid-sensing receptors, we also tested CRF and urocortin II when the same dose was dissolved in acidified vehicle, however, these injections also failed to induce hyperalgesia 8 h after their injection i.t. (data not shown). The lowest effective dose of urocortin I (1 μg i.t.) when administered intrathecally produced no effect on either von Frey fiber sensitivity or grip force responses after systemic (s.c.) administration (von Frey: urocortin I=2.6±0.6 responses, vehicle=3.1±0.7 responses ( t 18 =0.55, P =0.59); grip force: urocortin I=172.1±5.3 g, vehicle=175.0±0.6 g ( t 18 =0.40, P =0.69)). This suggests that the possible leakage of urocortin I from the spinal cord area to the periphery cannot account for its hyperalgesic effect when injected i.t. It was not until the dose of urocortin I was increased to 10 μg (in 200 μl) that it produced tactile mechanical hyperalgesia when injected peripherally (von Frey fiber assay), peaking 8 h after a s.c. ( Fig. 1D ; t 40 =4.82, P <0.0001) or i.p. injection ( Fig. 1D ; t 39 =2.92, P =0.006). Peripheral administration of urocortin I by the same high dose and routes had no effect on responses measured using the grip force assay ( Fig. 1H ; s.c. t 40 =0.85, P =0.40; i.p. t 39 =0.97, P =0.34). To further confirm the proposed site of action, we injected mice with 10 μg/5 μl i.t. urocortin I 30 min after a s.c. injection of 10 μg/200 μl astressin, a dose that is six times higher than that necessary to antagonize the effect of CRF injected intraperitoneally ( Martinez et al., 2002 ). Astressin failed to prevent the urocortin I-induced increases in either musculoskeletal sensitivity at 4 h (astressin plus urocortin I=164.2±6.4 responses verses urocortin I alone=161.7±4.0 responses ( t 20 =0.33, P =0.74)) or at 8 h (astressin plus urocortin I=161.9±6.0 responses verses urocortin I alone=164.7±5.2 responses ( t 20 =0.35, P =0.73)), or tactile sensitivity at 4 h (astressin plus urocortin I=9.8±2.2 g verses urocortin I alone=9.2±2.1 g ( t 20 =0.20, P =0.84)) or at 8 h (astressin plus urocortin 1=10.5±2.5 g verses urocortin I alone=8.6±2.2 g ( t 20 =0.54, P =0.59)), confirming that it is a central rather than peripheral CRF receptor population that mediates hyperalgesia. The long-term effect of urocortin I administered i.t. on von Frey fiber sensitivity was dose-related. Initially, when mice were injected with 1, 3 or 10 μg/5 μl urocortin I, peak effects measured 4–8 h after injection did not differ by dose, but higher doses elicited longer lasting and dose-related increases in von Frey fiber sensitivity, as shown at 24 after injection ( Fig. 2A : F 3,110 =11.58, P <0.0001). To confirm that the effect of urocortin I on von Frey fiber responses reflects hyperalgesia, we assessed the effect of morphine on these responses. Morphine, a potent analgesic in the category of a relatively non-selective opioid receptor agonist, was administered at doses of 3, 10 or 30 mg/kg i.p. to mice who had previously (6 h) been injected with urocortin I (10 μg in 5 μl i.t.) or vehicle. Thirty min after its injection, morphine inhibited urocortin I-induced von Frey fiber responses in a dose-related fashion ( Fig. 2B ; F 4,137 =8.07, P <0.0001). The effect of urocortin I administered i.t. on grip force responses was also dose-related ( Fig. 2C ). When mice were injected with 1, 3 or 10 μg/5 μl urocortin I, peak effects were again produced 4–8 h after injection, as shown at 4 h ( F 3,203 =9.46, P <0.0001) and at 8 h ( F 3,203 =4.44, P =0.005). To confirm that the effect of urocortin I on grip force responses reflects hyperalgesia rather than weakness, we again assessed the effect of morphine on these responses. Morphine was administered at doses of 3, 10 and 30 mg/kg i.p. to mice who had previously (6 h) been injected with urocortin I (10 μg in 5 μl i.t.) or vehicle. Thirty min after its injection, 30 mg/kg of morphine blocked the inhibitory effect of urocortin I on grip force responses ( Fig. 2D : F 4,137 =8.071, P =0.0001). It is noteworthy that the intrathecal injection of vehicle alone consistently produced a decrease in grip force responses ( t 99 =9.16, P <0.0001) as well as a transient increase in von Frey fiber responses ( t 56 =5.11, P <0.0001) 8 h later, consistent with a high sensitivity of these two assays to the stress of even a vehicle injection. When urocortin I was injected i.t. daily for 5 days at the lowest effective dose tested (1 μg/5 μl), or for 21 days at 3–4 day intervals with the higher dose (10 μg/5 μl), the maximum effect on von Frey responses observed 8 h after administration of urocortin I were each greater than that of vehicle in terms of area under the curve ( F 2,38 =6.11, P =0.005) but, while also significantly different than their pre-injection control values, did not differ between doses or from one injection to the next at either 1 μg ( F 5,12,60 =14.30, P <0.0001) or the 10 μg dose ( F 3,7,21 =7.54, P =0.001), as illustrated in figure 3A . In contrast, when measured 24 h after its injection, the time-interval most sensitive to its dose-related effects, the number of von Frey fiber responses after each injection of 1 μg of urocortin I increased over time ( Fig. 3B ; F 5,12,60 =8.93, P <0.0001). This indicates no tolerance to the hyperalgesic effect of urocortin I as a longer rather than a shorter duration of action was produced by each injection. In rats, CRF2 receptor activity is known to increase body temperature while CRF1 receptor activity does not ( Telegdy et al. , 2006 ). In contrast, injections of urocortin I directly into the dorsal raphe nucleus of mice cause hypothermia ( Turek & Ryabinin, 2005 ). In our study, urocortin I decreased body temperature 4 h (peak effect) after its central injection i.t. ( Fig. 4A : F 3,51 =7.03, P =0.0005) or i.c.v. ( Fig. 4B : F 3,38 =8.34, P =0.03) and this effect was mimicked to a much lesser extent by urocortin II, but only when injected i.t ( Fig. 4A ). CRF2 activity also decreases food intake in rats ( Yakabi et al. , 2011 ). Consistent with this, urocortin I decreased body weight 24 h after its injection i.t. ( Fig. 4E ; F 3,51 =7.03, P =0.0005) or i.c.v. ( Fig. 4F : F 3,38 =8.36, P =0.0002). This effect was also mimicked by urocortin II, but only when injected i.c.v. ( Fig. 4F ). CRF had no effect on body weight or body temperature when injected by either an i.t. or an i.c.v. route. However, 1 h after injection of CRF (10 μg) in a volume that distributes to both brain and spinal cord (20 μl i.t.), CRF increased the concentration of circulating corticosterone (513.9±24.66 ng/ml, n=8) when compared to vehicle (166.92±49.1 ng/ml, n=7, t 13 =6.59, P <0.0001) confirming the activity of this dose of CFR at CRF1 sites along the hypothalamic-pituitary-adrenal pathway. Injected parenterally, 10 μg of urocortin I decreased body temperature 4 h after its injection s.c. or i.p. when compared to vehicle-injected controls ( Fig. 4C : t 81 =16.29, P <0.0001), Fig 4D : t 39 =5.91, P <0.0001) suggesting that this dose was sufficient to induce a CRF2-mediated effect. In contrast to its effect on body temperature, 10 μg of urocortin I decreased body weight 24 h after a s.c. ( Fig. 4G : t 81 =10.27, P <0.0001) but not an i.p. injection ( Fig. 4H : t 39 =0.53, P =0.60). Together, these data confirm that these compounds are biologically active at both CRF1 and CRF2 sites in the CNS in the murine model. To determine whether the hyperalgesic effects of urocortin I were elicited by activation of CRF receptors, we tested the effect of a CRF1 receptor antagonist (NBI-35965 at 50 μg/5 μL i.t.) and of a CRF2 receptor antagonist (astressin 2B at 10 μg/5 μL i.t.), on urocortin I-induced hyperalgesia (3 μg/5 μl i.t.), measured using von Frey fiber sensitivity. When injected i.t. 15 min before 10 μg of urocortin I, NBI-35965 significantly attenuated the increase in von Frey fiber peak responses observed after the injection of urocortin I alone ( Fig. 5A : F 2,28 =16.16, P <0.0001). In contrast, astressin 2B failed to affect peak von Frey fiber responses detected 4–8 h after urocortin I ( Fig. 5B : F 2,28 =10.75, P =0.0003), indicating that the hyperalgesia measured using von Frey fiber sensitivity is mediated by CRF1 receptors. The effect of these antagonists was validated by their ability to influence urocortin I-induced decreases in body temperature (a CRF2 receptor effect) and their ability to attenuate increases in glucocorticoid release (a CRF1 mediated effect). The dose of astressin 2B used was sufficient to prevent the decrease in body temperature induced by 10 μg of urocortin I 4 h after injection (peak effect on temperature) ( Fig. 5H : F 2,28 =45, P =0.003), whereas NBI-35965 had no effect on urocortin I-induced hypothermia ( Fig. 5G : F 2,28 =7.25, P =0.003). This is consistent with a CRF2 receptor-mediated hypothermic effect of urocortins on body temperature. We then tested the effect of a CRF1 receptor antagonist (NBI-35965) and of a CRF2 receptor antagonist (astressin 2B) on urocortin I-induced hyperalgesia (3 μg/5 μl i.t.), measured using grip force. When injected i.t. 15 min before 10 μg of urocortin I, neither NBI-35965 nor astressin 2B attenuated the decrease in grip force responses observed after the injection of urocortin I alone ( Fig. 5D ; F 2,28 =4.43, P =0.021 and Fig. 5E ; F 2,28 =3.56, P =0.04). Although the selective CRF antagonists had no significant inhibitory effect on this hyperalgesia, each prevented the statistical significance of the decrease in grip force produced by urocortin I. Because each selective antagonist appeared to be only partially effective in preventing urocortin I-induced hyperalgesia measured using grip force, we then tested the effect of astressin, a non-selective CRF receptor antagonist (10 μg/10 μl i.t), on urocortin I-induced hyperalgesia (3 μg/5 μl i.t.). Higher doses of astressin 2B were not used due to observable behavioral effects that we determined might interfere with nociceptive responses. The hyperalgesic effect of urocortin I in the grip force assay was significantly inhibited by pretreatment with astressin delivered 30 min before urocortin I ( Fig. 5F : F 2,60 =6.48, P =0.003). The dose of astressin used was sufficiently low that it had no effect on CRF1-mediated hyperalgesia measured using von Frey fibers ( Fig. 5C : F 2,60 =8.03, P =0.0008) or on CRF2-mediated hypothermia ( Fig. 5 I : F 2,60 =42.69, P <0.0001). Thus, while selective inhibition of either CRF1 or CRF2 receptors has no effect on urocortin I-induced decreases in grip force, inhibition by a nonselective antagonist potently inhibited this same parameter. To determine the potential synergistic relationship between these two receptors, we explored the effect of both selective antagonists injected together with urocortin I (in series via the same delivery cannula) and found that they produced no greater antagonistic effect on von Frey fiber responses than either antagonist alone (data not shown), indicating a lack of synergy between these two receptors. Rather, the antihyperalgesic action of astressin may by due to an action at other CRF receptors than the CRF1 or CRF2 site.

Fibromyalgia is a painful condition in which the concentration of CRF is elevated in the cerebrospinal fluid ( McLean et al. , 2006 ). The pain of fibromyalgia is exacerbated by stress, which raises the possibility that stress hormones modulate nociception. Our studies show that urocortin I, a stress hormone with both CRF1 and CRF2 affinity, produces a persistent mechanical hyperalgesia when injected intrathecally in the mouse. This hyperalgesia is widespread, dose-dependent and attenuated by CRF receptor antagonists. No tolerance develops to the hyperalgesic effects of urocortin I, consistent with its potential to exacerbate chronic pain conditions. CRF and its receptors along with urocortins I, II and III are distributed in pain-relevant sites in the brain, spinal cord ( Korosi et al. , 2007 ) and periphery ( Martinez et al. , 2004b ) where CRF and its congeners modulate many types of nociception ( Lariviere & Melzack, 2000 ; Imbe et al. , 2006 ). Stress hormones alter pain in a fashion that depends of modality. Whether stress induces antinociception ( Lariviere & Melzack, 2000 ) or hyperalgesia ( Imbe et al. , 2006 ) depends on many factors, including the type, length and severity of the stress. The CRF receptor type and distribution relative to each modality of pain contributes to the dichotomous effects of these hormones ( Dautzenberg & Hauger, 2002 ; Ji & Neugebauer, 2008 ). For example, when injected i.t in mice, CRF produces no effect on thermal nociception (tail flick assay: ( Ayesta & Nikolarakis, 1989 ; Poree et al. , 1989 ; Song & Takemori, 1991 )), consistent with the failure of CRF overexpression to influence thermal hyperalgesia ( van Gaalen et al. , 2002 ). However, identical injections of CRF are antinociceptive in assays of chemical nociception (abdominal stretch), an effect thought to be mediated by dynorphin and kappa opioid receptors ( Przewlocki et al. , 1983 ; Nakazawa et al. , 1985 ; Song & Takemori, 1990 ). Even different doses of the same compound (e.g. CRF) can produce opposite effects due to differential affinities at CRF receptors ( Nijsen et al. , 2005 ). CRF is pronociceptive in models of abdominal pain ( Martinez & Tache, 2006 ) while urocortins exert abdominal antinociceptive effects in mice ( Martinez et al. , 2004a ) and rats ( Million et al. , 2006 ). Based on the conflicting effects of CRF receptor activity on various modalities of nociception, their effects specifically on musculoskeletal and tactile nociception cannot be extrapolated from those previously reported. To address this, the current study examined the influence of these compounds and their routes of administration specifically on musculoskeletal and tactile nociception. Our studies demonstrate that intrathecal injections of urocortin I produce mechanical hyperalgesia in both the von Frey fiber and grip force assays. Hyperalgesia produced by i.t. injections of urocortin I are consistent with the presence of neuronal projections containing CRF and urocortin I throughout the mouse spinal cord. These proteins are dense in the intermediate zone and surrounding the central canal ( Korosi et al. , 2007 ) but also present in lamina I and II of the dorsal horn ( Bittencourt et al. , 1999 ; Korosi et al. , 2007 ; Kim et al. , 2011 ). CRF1 and CRF2 receptor mRNA is also expressed throughout the spinal gray matter where CRF2 receptor mRNA exceeds that of CRF1 in the thoracic and lumbar spinal cord ( Korosi et al. , 2007 ; Kim et al. , 2011 ). Small dorsal root ganglia cells contain CRF protein together with only weak CRF2 receptor mRNA ( Kim et al. , 2011 ), but their expression increases in a rat model of neuropathic pain ( Kim et al. , 2011 ). This suggests a greater modulatory capacity of stress along damaged nociceptive pathways. The hyperalgesic effect of urocortin I on grip force responses and von Frey fiber sensitivity was reversed by morphine in a dose-related fashion, consistent with an effect of urocortin I on nociception. The relatively high dose of morphine required to antagonize urocortin I-induced hyperalgesia (30 mg/kg i.p.) usually suggests that mu opioid receptors, for which morphine has a relatively high affinity, do not mediate this antinociception. In fact, CRF receptors are known to antagonize the antinociceptive effect of morphine on thermal nociception ( Ayesta & Nikolarakis, 1989 ; Poree et al. , 1989 ; Song & Takemori, 1991 ). A similar antagonism may account for the attenuated action of morphine on urocortin I-induced mechanical hyperalgesia. Grip force was only decreased by urocortin I when injected i.t. In contrast, von Frey fiber sensitivity was increased by all routes except when injected i.c.v., indicating that the location of receptors capable of inducing tactile hyperalgesia appear to be more widespread than those influencing grip force. However, the lowest effective dose of urocortin I (1 μg) only increased von Frey fiber sensitivity when injected i.t., suggesting that nociceptive processing in the spinal cord area is a more sensitive target for urocortin I than the peripheral receptor populations present in skin, skeletal muscle and adipose tissue. In addition, the inability of peripherally injected antagonists to inhibit the hyperalgesic effects of urocortin I injected i.t. suggest that leakage of drug from the spinal area to the periphery does not likely account for its hyperalgesic action. Both assays of mechanical hyperalgesia appear to involve a CRF receptor, however, the CRF receptor population for grip force appears to differ from that involved in von Frey fiber sensitivity. Tactile hyperalgesia (von Frey) was inhibited by NBI-35962, a CRF1 receptor antagonist. This is consistent with previous reports that CRF1 receptors mediate tactile pain in rats (Hummel et al 2010). In contrast, the musculoskeletal hyperalgesic effect of urocortin I (grip force) was inhibited by astressin, the non-selective antagonist, and prevented by astressin 2B, the CRF2 receptor antagonist. This is consistent with the contribution of CRF2 receptors to the mediation of swim stress-induced musculoskeletal hyperalgesia in mice ( Abdelhamid et al. , 2013 ). Although astressin 2B prevented the development of musculoskeletal hyperalgesia, its failure to completely block the decrease in grip force induced by urocortin may be due to a shorter time-course of action or a lower efficacy in spite of higher selectively for the CRF2 receptor. Thus, selective inhibition of either CRF 1 or CRF 2 had no significant effect on urocortin 1-induced decrease in grip force while a nonselective antagonist did. This may indicate that urocortin I also influences hyperalgesia via an as yet undescribed CRF receptor that confounds the interpretation of our results. Repeated administration of CRF intrathecally leads to tolerance to its antinociceptive effect ( Song & Takemori, 1990 ; 1991 ), but in the present study no tolerance developed to the peak (8 h) mechanical hyperalgesia following injections of urocortin I at 3–4 day intervals. Rather, the magnitude of hyperalgesia 24 h after injection increased resulting in longer hyperalgesic episodes. This enduring pronociceptive action of urocortin I may contribute to the persistent hyperalgesia in response to repeated stress in rodents ( Abdelhamid et al. , 2013 ) as well as in many painful conditions in humans. It is somewhat perplexing that urocortin I-induced hyperalgesia was not mimicked by either CRF or urocortin II. The failure of CRF and urocortin II to produce hyperalgesia cannot be attributed to an insufficient dose, as these injections were biologically active. For example hypothermia, known to be induced by activation of CRF2 sites ( Turek & Ryabinin, 2005 ), was evident after injection of urocortin I by all routes (i.c.v., i.t., s.c., i.p.) and to a lesser extent after centrally injected urocortin II (i.t.). This hypothermia was prevented by astressin 2B but not by NBI-35962, indicating a classic CRF2-mediated effect. In addition, urocortin I (i.c.v., i.t., s.c.) and urocortin II (i.c.v.) each decreased food intake, likely by interrupting feeding and digestion (CRF1 receptor) or by postponing appetite and digestion (CRF2 receptor) ( Zorrilla et al. , 2003 ). While CRF had no effect in our study on either body temperature or weight, the dose used increased corticosterone, indicating physiologic activity along the hypothalamic-pituitary-adrenal axis. It is possible that the distribution of exogenously administered urocortin I differs from that of CRF and urocortin II allowing it better access to the relevant CRF receptor population. Alternatively, the unique hyperalgesic effects of urocortin I may be, in part, due to its association with CRF binding proteins, molecules that have a higher affinity for urocortin I. This binding protein can either chaperone urocortin I to its target ( Westphal & Seasholtz, 2006 ) or inhibit CRF-mediated activation of CRF1 receptors ( Manuel et. al 2014 ). In contrast, urocortin II can displace CRF agonists from these binding proteins ( Jahn et al. , 2005 ). Fibromyalgia is a painful condition in which the concentration of CRF is elevated in the cerebrospinal fluid ( McLean et al. , 2006 ). The pain of fibromyalgia is exacerbated by stress. In addition, fibromyalgia is frequently comorbid with sensitizing disorders including Crohn’s disease (49%) ( Buskila et al. , 1999 ), ulcerative colitis (19%) ( Buskila et al. , 1999 ), irritable bowel syndrome (about 70%) ( Veale et al. , 1991 ; Sivri et al. , 1996 ; Barton et al. , 1999 ; Sperber et al. , 1999 ), interstitial cystitis ( Alagiri et al. , 1997 ; Clauw et al. , 1997 ), and endometriosis ( Nothnick, 2004 ). In conditions like ulcerative colitis, immunoreactivity for urocortin I is increased and correlates with the severity of inflammation in colonic mucosa ( Saruta et al. , 2004 ). While urocortin I is antinociceptive in viscera, the present study predicts that urocortin I originating from the skin or gut might simultaneously cause tactile mechanical hyperalgesia similar to that after a peripheral injection. Furthermore, if the expression of CRF2 receptors increases in these conditions, as it does in a rat model of neuropathic pain ( Kim et al. , 2011 ), the mechanical hyperalgesia induced by endogenously released urocortin I might be enhanced even more than in healthy individuals. Now that the ability of urocortin I to induce hyperalgesia is established by the present study and the importance of CRF2 receptors in stress-induced musculoskeletal nociception known ( Abdelhamid et al. , 2013 ), further characterization of these CRF receptor populations is warranted in future studies. This model may elucidate the roles of CRF receptors in stress-induced mechanical hyperalgesia and serve as a screen for drug targets in painful conditions that are exacerbated by stress.