- Email: [email protected]
Behavioral responses to stress are intact in CRF-deficient mice
Brain Research 845 Ž1999. 14–20 www.elsevier.comrlocaterbres
Research report
Behavioral responses to stress are intact in CRF-deficient mice Adrian J. Dunn ) , Artur H. Swiergiel Department of Pharmacology and Therapeutics, Louisiana State UniÕersity Medical Center, ShreÕeport, LA 71130-3932, USA Accepted 27 July 1999
Abstract Corticotropin-releasing factor ŽCRF. has been implicated in endocrine and behavioral responses associated with stress. We have now studied the behavior of mice lacking the CRF gene ŽCRFko., comparing them to wild-type ŽWT. mice. Behaviors were observed in untreated mice, as well as following restraint or intraperitoneal administration of mouse interleukin-1b ŽmIL-1b .. In the multicompartment chamber ŽMCC., the behaviors of CRFko and WT mice were very similar, and prior restraint and IL-1b induced similar decreases in stimulus-contact times in both genotypes. In the elevated plus maze ŽEPM., restraint decreased the number of open arm entries but the behavior of both genotypes was very similar. In the open field ŽOF., the changes in locomotor activity in response to restraint were similar in both genotypes, although CRFko mice displayed slightly increased locomotor activity compared to WT mice. In both the MCC and the EPM, grooming behavior was increased by restraint, and was higher in the CRFko than in the WT mice. Compared to WT mice, CRFko mice had lower basal plasma concentrations of corticosterone which did not increase significantly following footshock. Thus, CRFko mice showed a clear dichotomy; the stress-related activation of the hypothalamo-pituitary-adrenal ŽHPA. axis was absent, whereas the stress-related behavioral responses thought to be mediated by brain CRF were unaffected. These results suggest that when mice develop in the absence of CRF, another factor Žor factors. assumes the behavioral functions normally ascribed to brain CRF, but not activation of the HPA axis. Alternatively, the natural modulator of behavior may not be CRF, but some other molecule that can act on receptors sensitive to CRF. Thus, redundant CNS mechanisms appear to be involved in stress-related behaviors. q 1999 Elsevier Science B.V. All rights reserved. Keywords: CRF; CRF-knockout; Stress; Behavior; Interleukin-1; Corticosterone
1. Introduction Corticotropin-releasing factor ŽCRF. was originally identified as a critical hypophysiotropic factor that could mediate the activation of the hypothalamo-pituitaryadrenocortical ŽHPA. axis in response to environmental stimuli w29x. CRF-containing cell bodies in the hypothalamic paraventricular nucleus project to the median eminence region of the hypothalamus where it is released and transported to the anterior pituitary from which it elicits secretion of ACTH. However, neurons containing CRF, mRNA for CRF, and various subtypes of CRF receptors have been identified outside the hypothalamus throughout much of the brain w6,21,22x. Stressful stimuli have been shown to increase the apparent release of CRF in the central amygdaloid nucleus as determined by in vivo microdialysis w17x. ) Corresponding [email protected]
author.
Fax:
q 1-318-675-7857;
e-mail:
Intracerebroventricular Ži.c.v.. administration of CRF induces a variety of behavioral patterns that resemble those observed during stress w10,15,20x. I.c.v. CRF administration also induces other physiological indices of stress, such as endocrine responses, changes in sympathetic and adrenomedullary activity, and activation of cerebral noradrenergic and dopaminergic but not serotonergic systems Žfor reviews, see Refs. w10,20x.. The concept that brain CRF or its receptors may mediate some of these responses has been bolstered by the repeated observation that administration of CRF receptor antagonists can attenuate or reverse many of the behavioral and endocrine responses observed during stress, but not the neurochemical ones w10,20x. Also, transgenic mice with hypersecretion of CRF display behavior suggestive of increased anxiety w14,26x. It was of interest to observe the behavioral and endocrine responses to restraint of mice lacking the ability to synthesize CRF to test the hypothesis that CRF is an essential mediator of physiological responses in stress. We also tested the responses to peripheral administration of
0006-8993r99r$ - see front matter q 1999 Elsevier Science B.V. All rights reserved. PII: S 0 0 0 6 - 8 9 9 3 Ž 9 9 . 0 1 9 1 2 - 5
A.J. Dunn, A.H. Swiergielr Brain Research 845 (1999) 14–20
interleukin-1b ŽIL-1b . which has been shown to mimic behavioral responses commonly observed during stress w9,25x. 2. Methods 2.1. Animals and materials Adult male mice were provided by Drs. Joseph A. Majzoub and Louis Muglia of the Children’s Hospital in Boston, MA. Wild type controls ŽWT. were homozygotes Ž Crfqrq . and the CRF knockout mice ŽCRFko. were transgenic homozygotes Ž crfyry . inserted with an inactivated CRF gene produced as described previously w19x. Two separate batches of mice of each genotype were shipped to Shreveport. The mice were separated by genotype, but the identity of the genotype was not known by the experimenters until completion of the experiments. The animals were housed singly in an AAALAC-accredited facility under a 12–12 light–dark cycle Žlights on at 0700 h. with free access to food and water. All procedures were approved by the Louisiana State University Medical Center Animal Care and Use Committee. Recombinant mouse IL-1b ŽmIL-1b . was obtained from R & D Systems, ŽMinneapolis, MN. and CRF was a gift from Dr. Jean Rivier ŽThe Peptide Laboratory, The Salk Institute, San Diego, CA.. 2.2. Multicompartment chamber (MCC) The MCC apparatus and testing procedures have been described previously w2,3x. In brief, the apparatus consisted of a box divided into nine interconnecting compartments, each containing a hole in the floor. Recessed below each hole, a wire ball was rigidly attached to serve as a stimulus. A stimulus contact was defined when a mouse made any contact with this wire ball. The test was initiated by placing the mouse in the central compartment of the apparatus and the frequency and duration of contacts, and the number of compartment entries as well as grooming and rears were scored for 25 min.
15
a box 59 = 59 cm with its floor divided into four squares and illuminated by white light. Locomotor activity was scored as line crossings and rears for 5 min. A line crossing was scored when all four paws left one square and entered another. Rears were scored when a mouse raised both its front paws from the floor. 2.5. Experimental procedure and behaÕioral obserÕations The MCC and OF tests were conducted in a room separate from the animal colony room, between 0900 and 0500 AM. The EPM test was performed between 0900 and 1100 h under red light after habituation to the room for 3 h. Animals were restrained by placing them for 30 min in 50 ml centrifuge tubes as previously described, and tested immediately after removal from the tubes w3x. Following each test, the apparati were cleaned with 1% acetic acid. mIL-1b was administered intraperitoneally Ži.p.. at a dose of 100 ngrmouse 90 min before placement in the MCC. Footshock was administered through the grid floor of a small Plexiglas box using a random shock generator, such that 30 1-s shocks of 0.2 mA were delivered in the 30-min period w8x. 2.6. Assay of plasma corticosterone Blood was collected from the retro-ocular artery or following decapitation. Plasma was separated by centrifugation, and stored frozen. Corticosterone was assayed using w125 Ixradioimmunoassay kits supplied by ICN Biomedicals ŽCosta Mesa, CA. as described by the manufacturer. 2.7. Statistical analyses Data were analyzed by multifactorial analysis of variance followed by pairwise comparisons using Fisher’s Least Significant Difference test when appropriate. Data presented are means " S.E.M.
3. Results
2.3. EleÕated plus maze (EPM) The EPM was made of black Plexiglas with 5 cm wide arms elevated 40 cm above the floor w13x. Two opposite arms were enclosed by transparent walls Ž30 cm long = 5 cm wide = 15 cm high. and two arms were open Ž30 cm long = 5 cm wide.. The mouse was placed in the center of the apparatus facing an enclosed arm and observed for 5 min. The time spent in the enclosed and open arms and the total number of entries into the enclosed and open arms was recorded. 2.4. Open field (OF) Spontaneous locomotor activity was quantified in an OF immediately after completion of the EPM test. The OF was
3.1. BehaÕioral obserÕations in the MCC Mice were first tested in the MCC with or without prior restraint. Consistent with our earlier studies w3x, restraint induced a statistically significant Ž F1,30 s 72, p - 0.0001. reduction in the mean time per contact ŽFig. 1.. There were no significant differences between WT and CRFko mice Ž F1,30 s 1.37., but there was a significant restraint= genotype interaction Ž F1,30 s 6.0, p - 0.05.. A significant interaction was not observed in an earlier experiment, and the effect in this experiment probably reflects the decreased stimulus-contact time in the quiet CRFko mice, which was not itself statistically significant. Restraint significantly decreased the number of compartment entries
16
A.J. Dunn, A.H. Swiergielr Brain Research 845 (1999) 14–20
tion Ž F1,29 s 0.. The total number of are entries, a useful indicator of locomotor activity was not affected by genotype ŽWT 26.8 " 5.2; CRFko 27.3 " 6.0, F1,29 s 0.3., but was significantly decreased by restraint ŽWT 17.7 " 7.2; CRFko 19.3 " 5.0, F1,29 s 17.4, p - 0.001. and there was no significant restraint= genotype interaction Ž F1,29 s 0.06.. 3.3. BehaÕioral obserÕations in the OF The mice were observed in the OF immediately after completion of an EPM test. ANOVA revealed no significant effects of restraint Žcrossings: F1,36 s 2.7; rears: 1.39. or genotype Ž F1,36 s 1.40 and 1.33. and no significant restraint= genotype interaction Ž F1,36 s 0.58 and 0.26. on the number of crossings and rears in the OF in the first trial. In a second test, the CRFko mice displayed more line crossings than WT mice Ž F1,36 s 4.6, p - 0.05.. There were no other significant differences between the genotypes and treatments. When the data from both trials were pooled ŽFig. 4., CRFko mice displayed more locomotor
Fig. 1. Effect of restraint on the mean stimulus-contact time and the number of compartment entries in the MCC. Wild type and CRF-knockout mice were restrained for 30 min and immediately placed in the MCC and observed for 25 min. UUU Significantly different from unrestrained mice ŽUU p- 0.01, UUU p- 0.001, N s8–9..
Ž F1,30 s 4.36, p - 0.05., but there was no statistically significant difference between the two strains Ž F1,30 s 1.11. and no significant restraint= genotype interaction Ž F1,30 s 1.35.. Also in the MCC, administration of mIL-1b Ž100 ng i.p.. significantly reduced the mean time per contact ŽFig. 2; F1,36 s 50, p - 0.0001.. However, there was no effect of genotype Ž F1,36 s 0.47. and no mIL-1b = genotype interaction Ž F1,36 s 2.1.. In this experiment, none of the other behaviors scored was significantly affected by the genotype or the mIL-1b. 3.2. BehaÕioral obserÕations in the EPM Mice were tested in the EPM the day after a session in the MCC ŽFig. 3.. The number of entries into the open arms was decreased by restraint Ž F1,29 s 14.2, p - 0.001., but there was no effect of genotype Ž F1,29 s 0.3., nor a restraint = genotype interaction Ž F1,29 s 0.. Restraint marginally decreased the time spent on the open arms Ž F1,29 s 4.02, p s 0.054., but there was no significant restraint= genotype interaction Ž F1,29 s 0.02., nor any effect of genotype Ž F1,29 s 0.34.. The number of entries into the enclosed arms was also decreased by restraint Ž F1,29 s 9.0, p - 0.01., but there was no significant effect of genotype Ž F1,29 s 0.0., nor a restraint= genotype interac-
Fig. 2. Effect of mIL-1b on mean stimulus-contact time and the number of compartment entries in the MCC. WT and CRFko mice were injected i.p. with mIL-1b Ž100 ngrmouse. or saline and 90 min later placed in the MCC for 25 min. Significantly different from saline-injected mice ŽUUU p - 0.001, N s10..
A.J. Dunn, A.H. Swiergielr Brain Research 845 (1999) 14–20
17
straint = genotype interactions Ž F1,30 s 0.04 and F1,30 s 1.04.. Grooming was also increased by restraint in the EPM. The total duration increased ŽFig. 5; F1,29 s 7.2, p - 0.05., and the number of bouts was increased Ž F1,29 s 7.1, p - 0.05.. In both cases, there was a significant effect of genotype such that CRFko mice displayed more grooming than WT Ž F1,29 s 11.0, p - 0.01 and F1,29 s 12.9, p - 0.01., but there were no statistically significant restraint = genotype interactions Ž F1,29 s 0.3 and 1.2.. Grooming was not scored in the OF. 3.5. Plasma corticosterone Plasma concentrations of corticosterone were measured following 20 min footshock, and again 2 h after i.p. injection of 100 ng mIL-1b. CRFko mice had substantially lower concentrations of plasma corticosterone compared to WT mice. Footshock elicited the expected increase in WT mice, but no increase was observed in the CRFko mice ŽFig. 6.. ANOVA indicated a statistically significant geno-
Fig. 3. Effect of restraint on behavior in the EPM. WT and CRFko mice were restrained for 30 min and immediately placed in the EPM for 5 min. Top graph: number of open arm entries; Middle graph: time spent on the open arms; Bottom graph: number of enclosed arm entries. U Significantly different from unrestrained mice ŽU p- 0.05, UU p- 0.01, N s8–9..
activity in the OF compared to WT mice, a slightly higher number of line crossings Ž F1,72 s 5.9, p - 0.05. and rears Ž F1,72 s 3.1, p - 0.08., but no genotype= restraint interaction Ž F1,72 s 0.0 and 1.44; crossings and rears, respectively.. 3.4. Grooming behaÕior Restraint increased grooming in the MCC; both the total duration ŽFig. 5; F1,30 s 14.2, p - 0.001. and the frequency Žthe number of grooming bouts; F1,30 s 10.2, p 0.01.. The total duration of grooming was significantly elevated in the CRFko compared to the WT mice Ž F1,30 s 4.52, p - 0.05., but there was no effect of genotype on the frequency Ž F1,30 s 1.25.. There were no significant re-
Fig. 4. Effect of restraint on behavior in the OF. WT and CRFko mice were restrained for 30 min and tested in the EPM for 5 min, after which they were immediately placed in the OF for 5 min. Upper graph: the number of line crossings; lower graph: the number of rears. Data were pooled from two separate experiments. There was a significant effect of genotype Ž p- 0.05. on the number of line crossings Ž N s10..
18
A.J. Dunn, A.H. Swiergielr Brain Research 845 (1999) 14–20
4. Discussion
Fig. 5. Effect of restraint on grooming behavior in the MCC Župper graph. and EPM Žlower graph.. The total time spent grooming observed during the observation periods in the MCC Ž25 min. and the EPM Ž5 min.. Data taken from the experiments in Figs. 1 and 3. U Significantly different from unrestrained mice ŽU p- 0.05, N s8–9..
type = treatment interaction Ž F1,31 s 193, p - 0.0001., reflecting the lack of response in the CRFko mice. IL-1b treatment increased plasma corticosterone in the WT mice, but the response was dramatically reduced in the CRFko mice ŽFig. 6.. ANOVA indicated a statistically significant genotype= treatment interaction Ž F3,28 s 44, p - 0.0001.. The small increase induced by IL-1b in the CRFko mice was statistically significant Ž t 13 s 2.31, p - 0.05..
Behavior in the MCC has been shown to be reliably affected by a variety of stressors, such that the stimuluscontact time is reduced without consistent changes in locomotor activity in rats w1x and mice w3,5x. These responses were closely mimicked by i.c.v. administration of low doses of CRF in mice w3,5x and rats w24x. I.c.v. pretreatment with the CRF-receptor antagonist, a-helical CRF9 – 41 , prevented the restraint-induced changes in behavior w4x, suggesting that brain CRF plays an important role in the stress-related alterations of behavior in the MCC. Furthermore, administration of IL-1 either i.p. w9x or i.c.v. w25x elicited responses in the MCC very similar to those of restraint and i.c.v. CRF. These responses were also attenuated by i.c.v. administration of a-helical CRF9 – 41 w9,25x. The data from the present experiments are similar to those observed earlier in CD-1 mice. Similar behavioral scores were recorded, and there were similar responses to restraint. Mice that lacked the CRF gene, and hence presumably the capacity to synthesize CRF, showed behavioral responses to restraint in the MCC that very closely resembled those of wild type animals. These results confirm those of an earlier experiment in which the behavior in the MCC of CRFko mice was indistinguishable from that of WT w31x. In those experiments, we also observed very similar responses to i.c.v. administration of 20 ng CRF, suggesting that in CRFko mice there is no substantial alteration of the activity of receptors acted upon by i.c.v. CRF. A similar lack of difference between the genotypes was observed in the EPM. Consistent with previous observations w13x, restraint decreased open arm entries and the time spent on the open arms Žalthough the latter effect was only marginally statistically significant.. However, the behavior of the CRFko mice was very similar to that of the WT mice, consistent with results obtained in another labo-
Fig. 6. Effect of footshock and mIL-1b on plasma corticosterone. Blood was collected from WT and CRFko mice Ž n s 17. immediately after removal from their home cages, and then again on the following day immediately after the mice received footshocks for 30 min. Later, the mice were injected with 100 ng mIL-1b i.p. or saline and trunk blood collected following decapitation 2 h later. UUU Significantly different from quiet or saline-injected mice Ž p - 0.0001..
A.J. Dunn, A.H. Swiergielr Brain Research 845 (1999) 14–20
ratory w31x. The total number of entries into both the open and closed arms, a good measure of the overall locomotor activity, was almost identical in both genotypes, with or without restraint. Even though the effects of restraint or genotype might have been affected by the previous exposure of all animals to the EPM, the CRFko and WT mice displayed very similar behavioral response to restraint in the OF. The major distinction between the genotypes was a slightly higher number of line-crossings and rears in the CRFko mice, a response resembling that observed after immobilization, footshock or i.c.v. administration of CRF w16x. Together with the locomotor activity data from the MCC and EPM, these results strongly suggest that the general motor performance is not affected by the CRF deficiency. We observed a significant increase in grooming behavior in CRFko mice relative to WT in both the MCC and in the EPM. Increased grooming in the EPM was also observed in CRFko mice in another laboratory w31x. Grooming can be induced in rats and mice by i.c.v. CRF w11x or ACTH w12x. The grooming associated with the mild stress of novelty may be mediated by ACTH w12x. The enhanced response to CRF might be expected if the lack of CRF in CRFko mice were to result in a supersensitivity of CRF receptors. However, no difference in grooming was observed in the MCC in response to i.c.v. CRF ŽDunn and Swiergiel, unpublished observations.. It is possible that the increased grooming in the CRF-deficient mice is related to compensatory changes in other neuropeptide mediators of grooming, such as oxytocin, prolactin or b-endorphin w7x. The CRFko mice displayed responses to mIL-1b in the MCC that closely resembled those of WT mice, indicating that the ability of IL-1 to induce behavioral responses was not impaired, even though these responses in CD-1 mice appear to depend, at least in part, on functional CRF-receptors w9,25x. We also found that the anorectic response to IL-1 was unimpaired in CRFko mice w27x. Consistent with earlier observations in CRFko mice w18x, we observed low basal concentrations of plasma corticosterone. Footshock elevated plasma corticosterone in the WT mice, but there was no such increase in the CRFko mice. The response to IL-1b in the CRFko mice was substantially reduced, but was still statistically significant. A statistically significant increase of plasma corticosterone following IL-1b was observed in another experiment with CRFko mice, suggesting that this response is real. It is possible that catecholamines or some other peripheral factor was responsible for this elevation. Thus, in contrast to the very small changes observed in behavior, the HPA axis activation in CRFko mice was dramatically impaired. Mice lacking the gene for CRF-1 receptor exhibit decreased stress-related release of ACTH and corticosterone w23,28x. However, they also exhibit decreased anxiety in defensive withdrawal and the EPM w23x and in the light– dark box w28x. In the OF, small increases in locomotor
19
activity were noted w28x. These results are consistent with those reported here in that the stress-related behavioral responses were not absent. Thus, CRF and the CRF-1 receptor appear to be critical for the activation of ACTH and corticosterone secretion, but neither is essential for the behavioral responses. Apparently, other receptors Že.g., CRF-2 receptors. can assume some of the behavioral functions of CRF-1 receptors in the CNS. The general conclusion of these studies must be that CRF-knockout mice display behavior and behavioral responses to stressors remarkably similar to those of wild type animals. This conclusion runs counter to the inference from a large number of studies that brain CRF plays a major role in the behavioral responses normally observed during stress w10,20x. Because attenuation or prevention of the normal response during stress has been observed using a variety of different CRF-receptor antagonists and antibodies to CRF and its receptors, it is unlikely that these results are artifactual. The most obvious explanation is that mice that develop in the absence of CRF acquire other mechanisms to perform some of the normal functions of CRF in wild-type animals. Given that CRF receptors are present in CRFko mice, it is possible that another substance may assume some of the functions of CRF in WT animals. An obvious candidate is urocortin which is known to bind to the known CRF-receptors with affinities similar to those of CRF w30x. However, the anatomic distribution of urocortin does not correlate at all well with that of CRF-receptors w30,31x. Neither urotensin nor sauvagine, both of which bind CRF-receptors, has been detected in mice or rats. Perhaps some undiscovered peptide or nonpeptide ligandŽs. israre involved. Urocortin does not appear to substitute for CRF in the activation of the HPA axis, because, although it is a potent stimulator of ACTH secretion from the pituitary w30x, little, if any, stress-related elevation of plasma ACTH or corticosterone is observed in CRFko mice ŽRef. w18x and Fig. 6.. An alternative possibility is that CRF does not mediate the behavioral responses to stressors in WT mice. Perhaps, urocortin, or some other molecule, performs this function by acting on receptors that can also be activated by CRF. This would explain both the effects of i.c.v. CRF, and the ability of antagonists of presumed CRF receptors to attenuate the responses in stress. The most parsimonious conclusion is that redundancy exists in the CNS mechanisms of stress-related behaviors, although receptors that can be activated by CRF are involved. Tentatively we can conclude that there may be redundancy in both the ligands for CRF receptors, and in the receptors themselves.
Acknowledgements This research was supported by grants from the US National Institute of Mental Health ŽMH46261. and the US National Institute of Neurological Diseases and Stroke
20
A.J. Dunn, A.H. Swiergielr Brain Research 845 (1999) 14–20
ŽNS25370.. We thank Dr. Joseph A. Majzoub and Dr. Louis Muglia ŽChildren’s Hospital, Boston, MA. for providing the homozygous CRF knockout mice Ž crfyry . and the wild type Ž crfqrq . mice. The technical assistance of Michael Adamkiewicz and Galina Mikhaylova is greatly appreciated.
w17x
w18x
w19x
References w20x w1x A.F.T. Arnsten, C. Berridge, D.S. Segal, Stress produces opioid-like effects on investigatory behavior, Pharmacol. Biochem. Behav. 22 Ž1985. 803–809. w2x A.T. Arnsten, D.S. Segal, Naloxone alters locomotion and interaction with environmental stimuli, Life Sci. 25 Ž1979. 1035–1042. w3x C.W. Berridge, A.J. Dunn, Corticotropin-releasing factor elicits naloxone-sensitive stress-like alterations in exploratory behavior in mice, Regul. Pept. 16 Ž1986. 83–93. w4x C.W. Berridge, A.J. Dunn, A corticotropin-releasing factor antagonist reverses the stress-induced changes of exploratory behavior in mice, Horm. Behav. 21 Ž1987. 393–401. w5x C.W. Berridge, A.J. Dunn, Restraint-stress-induced changes in exploratory behavior appear to be mediated by norepinephrinestimulated release of CRF, J. Neurosci. 9 Ž1989. 3513–3521. w6x E.B. De Souza, T.R. Insel, M.H. Perrin, J. Rivier, W.W. Vale, M.J. Kuhar, Corticotropin-releasing factor receptors are widely distributed within the rat central nervous system: an autoradiographic study, J. Neurosci. 5 Ž1985. 3189–3203. w7x F. Drago, A. Contarino, L. Busa, ` The expression of neuropeptide-induced excessive grooming behavior in dopamine D1 and D 2 receptor-deficient mice, Eur. J. Pharmacol. 365 Ž1999. 125–131. w8x A.J. Dunn, Stress-related changes in cerebral catecholamine and indoleamine metabolism: lack of effect of adrenalectomy and corticosterone, J. Neurochem. 51 Ž1988. 406–412. w9x A.J. Dunn, M. Antoon, Y. Chapman, Reduction of exploratory behavior by intraperitoneal injection of interleukin-1 involves brain corticotropin-releasing factor, Brain Res. Bull. 26 Ž1991. 539–542. w10x A.J. Dunn, C.W. Berridge, Physiological and behavioral responses to corticotropin-releasing factor administration: is CRF a mediator of anxiety or stress responses?, Brain Res. Rev. 15 Ž1990. 71–100. w11x A.J. Dunn, C.W. Berridge, Y.I. Lai, T.L. Yachabach, CRF-induced excessive grooming behavior in rats and mice, Peptides 8 Ž1987. 841–844. w12x A.J. Dunn, E.J. Green, R.L. Isaacson, Intracerebral adrenocorticotropic hormone mediates novelty-induced grooming in the rat, Science 203 Ž1979. 281–283. w13x A.L. Gorman, A.J. Dunn, b-Adrenergic receptors are involved in stress-related behavioral changes, Pharmacol. Biochem. Behav. 45 Ž1993. 1–7. w14x S.C. Heinrichs, M.P. Stenzel-Poore, L.H. Gold, E. Battenberg, F.E. Bloom, G.F. Koob, W.W. Vale, E. Merlo Pich, Learning impairment in transgenic mice with central overexpression of corticotropin-releasing factor, Neurosci. 74 Ž1996. 303–311. w15x G.F. Koob, F.E. Bloom, Corticotropin-releasing factor and behavior, Fed. Proc. 44 Ž1985. 259–263. w16x E.H.Y. Lee, Y.P. Tang, C.Y. Chai, Stress and corticotropin-releasing
w21x
w22x
w23x
w24x
w25x
w26x
w27x
w28x
w29x
w30x
w31x
factor potentiate center region activity of mice in an open field, Psychopharmacology 93 Ž1987. 320–324. E.M. Merlo Pich, G.F. Koob, M. Heilig, F. Menzaghi, W. Vale, F. Weiss, Corticotropin-releasing factor release from the mediobasal hypothalamus of the rat as measured by microdialysis, Neuroscience 55 Ž1993. 695–707. L. Muglia, L. Jacobson, P. Dikkes, J.A. Majzoub, Corticotropin-releasing hormone deficiency reveals major fetal but not adult glucocorticoid need, Nature 373 Ž1995. 427–432. L. Muglia, L. Jacobson, J.A. Majzoub, Production of corticotropinreleasing hormone-deficient mice by targeted mutation in embryonic stem cells, Ann. N. Y. Acad. Sci. ŽU.S.A.. 780 Ž1996. 49–59. M.J. Owens, C.B. Nemeroff, Physiology and pharmacology of corticotropin-releasing factor, Pharmacol. Rev. 43 Ž1991. 425–474. E. Potter, S. Sutton, C. Donaldson, R. Chen, M. Perrin, K. Lewis, P.E. Sawchenko, W. Vale, Distribution of corticotropin-releasing factor receptor mRNA expression in the rat brain and pituitary, Proc. Natl. Acad. Sci. U.S.A. 91 Ž1994. 8777–8781. P.E. Sawchenko, L.W. Swanson, Organization of CRF immunoreactive cells and fibers in the rat brain: immunohistochemical studies, in: E.B. De Souza, C.B. Nemeroff ŽEds.., Corticotropin-Releasing Factor: Basic and Clinical Studies of a Neuropeptide, CRC Press, Boca Raton, FL, 1990, pp. 29–52. G.W. Smith, J.-M. Aubry, F. Dellu, A. Contarino, L.M. Bilezikjian, L.H. Gold, R. Chen, Y. Marchuk, C. Hauser, C.A. Bentley, P.E. Sawchenko, G.F. Koob, W. Vale, K.-F. Lee, Corticotropin releasing factor receptor 1-deficient mice display decreased anxiety, impaired stress response, and aberrant neuroendocrine development, Neuron 20 Ž1998. 1093–1102. F. Spadaro, C.W. Berridge, H.A. Baldwin, A.J. Dunn, Corticotropin-releasing factor acts via a third ventricle site to reduce exploratory behavior in rats, Pharmacol. Biochem. Behav. 36 Ž1990. 305–309. F. Spadaro, A.J. Dunn, Intracerebroventricular administration of interleukin-1 to mice alters investigation of stimuli in a novel environment, Brain Behav. Immunol. 4 Ž1990. 308–322. M.P. Stenzel-Poore, S.C. Heinrichs, S. Rivest, G.F. Koob, W.W. Vale, Overproduction of corticotropin-releasing factor in transgenic mice: a genetic model of anxiogenic behavior, J. Neurosci. 14 Ž1994. 2579–2584. A.H. Swiergiel, A.J. Dunn, CRF-deficient mice respond like wild type mice to hypophagic stimuli, Pharmacol. Biochem. Behav. 64 Ž1999. 59–64. P. Timpl, R. Spanagel, I. Sillaber, A. Kresse, J.M.H.M. Reul, G.K. Stalla, V. Blanquet, T. Steckler, F. Holsboer, W. Wurst, Impaired stress response and reduced anxiety in mice lacking a functional corticotropin-releasing hormone receptor 1, Nat. Genet. 19 Ž1998. 162–166. W. Vale, J. Spiess, C. Rivier, J. Rivier, Characterization of a 41-residue ovine hypothalamic peptide that stimulates secretion of corticotropin and b-endorphin, Science 213 Ž1981. 1394–1397. J. Vaughan, C. Donaldson, J. Bittencourt, M.H. Perrin, K. Lewis, S. Sutton, R. Chan, A.V. Turnbull, D. Lovejoy, C. Rivier, P.E. Sawchenko, W. Vale, Urocortin, a mammalian neuropeptide related to fish urotensin I and to corticotropin-releasing factor, Nature 378 Ž1995. 287–292. S.C. Weninger, A.J. Dunn, L.J. Muglia, P. Dikkes, K.A. Miczek, A.H. Swiergiel, C.W. Berridge, J.A. Majzoub, Stress-induced behaviors require the CRH receptor, but not CRH, Proc. Natl. Acad. Sci. U.S.A. 96 Ž1999. 8283–8288.
Research report
Behavioral responses to stress are intact in CRF-deficient mice Adrian J. Dunn ) , Artur H. Swiergiel Department of Pharmacology and Therapeutics, Louisiana State UniÕersity Medical Center, ShreÕeport, LA 71130-3932, USA Accepted 27 July 1999
Abstract Corticotropin-releasing factor ŽCRF. has been implicated in endocrine and behavioral responses associated with stress. We have now studied the behavior of mice lacking the CRF gene ŽCRFko., comparing them to wild-type ŽWT. mice. Behaviors were observed in untreated mice, as well as following restraint or intraperitoneal administration of mouse interleukin-1b ŽmIL-1b .. In the multicompartment chamber ŽMCC., the behaviors of CRFko and WT mice were very similar, and prior restraint and IL-1b induced similar decreases in stimulus-contact times in both genotypes. In the elevated plus maze ŽEPM., restraint decreased the number of open arm entries but the behavior of both genotypes was very similar. In the open field ŽOF., the changes in locomotor activity in response to restraint were similar in both genotypes, although CRFko mice displayed slightly increased locomotor activity compared to WT mice. In both the MCC and the EPM, grooming behavior was increased by restraint, and was higher in the CRFko than in the WT mice. Compared to WT mice, CRFko mice had lower basal plasma concentrations of corticosterone which did not increase significantly following footshock. Thus, CRFko mice showed a clear dichotomy; the stress-related activation of the hypothalamo-pituitary-adrenal ŽHPA. axis was absent, whereas the stress-related behavioral responses thought to be mediated by brain CRF were unaffected. These results suggest that when mice develop in the absence of CRF, another factor Žor factors. assumes the behavioral functions normally ascribed to brain CRF, but not activation of the HPA axis. Alternatively, the natural modulator of behavior may not be CRF, but some other molecule that can act on receptors sensitive to CRF. Thus, redundant CNS mechanisms appear to be involved in stress-related behaviors. q 1999 Elsevier Science B.V. All rights reserved. Keywords: CRF; CRF-knockout; Stress; Behavior; Interleukin-1; Corticosterone
1. Introduction Corticotropin-releasing factor ŽCRF. was originally identified as a critical hypophysiotropic factor that could mediate the activation of the hypothalamo-pituitaryadrenocortical ŽHPA. axis in response to environmental stimuli w29x. CRF-containing cell bodies in the hypothalamic paraventricular nucleus project to the median eminence region of the hypothalamus where it is released and transported to the anterior pituitary from which it elicits secretion of ACTH. However, neurons containing CRF, mRNA for CRF, and various subtypes of CRF receptors have been identified outside the hypothalamus throughout much of the brain w6,21,22x. Stressful stimuli have been shown to increase the apparent release of CRF in the central amygdaloid nucleus as determined by in vivo microdialysis w17x. ) Corresponding [email protected]
author.
Fax:
q 1-318-675-7857;
e-mail:
Intracerebroventricular Ži.c.v.. administration of CRF induces a variety of behavioral patterns that resemble those observed during stress w10,15,20x. I.c.v. CRF administration also induces other physiological indices of stress, such as endocrine responses, changes in sympathetic and adrenomedullary activity, and activation of cerebral noradrenergic and dopaminergic but not serotonergic systems Žfor reviews, see Refs. w10,20x.. The concept that brain CRF or its receptors may mediate some of these responses has been bolstered by the repeated observation that administration of CRF receptor antagonists can attenuate or reverse many of the behavioral and endocrine responses observed during stress, but not the neurochemical ones w10,20x. Also, transgenic mice with hypersecretion of CRF display behavior suggestive of increased anxiety w14,26x. It was of interest to observe the behavioral and endocrine responses to restraint of mice lacking the ability to synthesize CRF to test the hypothesis that CRF is an essential mediator of physiological responses in stress. We also tested the responses to peripheral administration of
0006-8993r99r$ - see front matter q 1999 Elsevier Science B.V. All rights reserved. PII: S 0 0 0 6 - 8 9 9 3 Ž 9 9 . 0 1 9 1 2 - 5
A.J. Dunn, A.H. Swiergielr Brain Research 845 (1999) 14–20
interleukin-1b ŽIL-1b . which has been shown to mimic behavioral responses commonly observed during stress w9,25x. 2. Methods 2.1. Animals and materials Adult male mice were provided by Drs. Joseph A. Majzoub and Louis Muglia of the Children’s Hospital in Boston, MA. Wild type controls ŽWT. were homozygotes Ž Crfqrq . and the CRF knockout mice ŽCRFko. were transgenic homozygotes Ž crfyry . inserted with an inactivated CRF gene produced as described previously w19x. Two separate batches of mice of each genotype were shipped to Shreveport. The mice were separated by genotype, but the identity of the genotype was not known by the experimenters until completion of the experiments. The animals were housed singly in an AAALAC-accredited facility under a 12–12 light–dark cycle Žlights on at 0700 h. with free access to food and water. All procedures were approved by the Louisiana State University Medical Center Animal Care and Use Committee. Recombinant mouse IL-1b ŽmIL-1b . was obtained from R & D Systems, ŽMinneapolis, MN. and CRF was a gift from Dr. Jean Rivier ŽThe Peptide Laboratory, The Salk Institute, San Diego, CA.. 2.2. Multicompartment chamber (MCC) The MCC apparatus and testing procedures have been described previously w2,3x. In brief, the apparatus consisted of a box divided into nine interconnecting compartments, each containing a hole in the floor. Recessed below each hole, a wire ball was rigidly attached to serve as a stimulus. A stimulus contact was defined when a mouse made any contact with this wire ball. The test was initiated by placing the mouse in the central compartment of the apparatus and the frequency and duration of contacts, and the number of compartment entries as well as grooming and rears were scored for 25 min.
15
a box 59 = 59 cm with its floor divided into four squares and illuminated by white light. Locomotor activity was scored as line crossings and rears for 5 min. A line crossing was scored when all four paws left one square and entered another. Rears were scored when a mouse raised both its front paws from the floor. 2.5. Experimental procedure and behaÕioral obserÕations The MCC and OF tests were conducted in a room separate from the animal colony room, between 0900 and 0500 AM. The EPM test was performed between 0900 and 1100 h under red light after habituation to the room for 3 h. Animals were restrained by placing them for 30 min in 50 ml centrifuge tubes as previously described, and tested immediately after removal from the tubes w3x. Following each test, the apparati were cleaned with 1% acetic acid. mIL-1b was administered intraperitoneally Ži.p.. at a dose of 100 ngrmouse 90 min before placement in the MCC. Footshock was administered through the grid floor of a small Plexiglas box using a random shock generator, such that 30 1-s shocks of 0.2 mA were delivered in the 30-min period w8x. 2.6. Assay of plasma corticosterone Blood was collected from the retro-ocular artery or following decapitation. Plasma was separated by centrifugation, and stored frozen. Corticosterone was assayed using w125 Ixradioimmunoassay kits supplied by ICN Biomedicals ŽCosta Mesa, CA. as described by the manufacturer. 2.7. Statistical analyses Data were analyzed by multifactorial analysis of variance followed by pairwise comparisons using Fisher’s Least Significant Difference test when appropriate. Data presented are means " S.E.M.
3. Results
2.3. EleÕated plus maze (EPM) The EPM was made of black Plexiglas with 5 cm wide arms elevated 40 cm above the floor w13x. Two opposite arms were enclosed by transparent walls Ž30 cm long = 5 cm wide = 15 cm high. and two arms were open Ž30 cm long = 5 cm wide.. The mouse was placed in the center of the apparatus facing an enclosed arm and observed for 5 min. The time spent in the enclosed and open arms and the total number of entries into the enclosed and open arms was recorded. 2.4. Open field (OF) Spontaneous locomotor activity was quantified in an OF immediately after completion of the EPM test. The OF was
3.1. BehaÕioral obserÕations in the MCC Mice were first tested in the MCC with or without prior restraint. Consistent with our earlier studies w3x, restraint induced a statistically significant Ž F1,30 s 72, p - 0.0001. reduction in the mean time per contact ŽFig. 1.. There were no significant differences between WT and CRFko mice Ž F1,30 s 1.37., but there was a significant restraint= genotype interaction Ž F1,30 s 6.0, p - 0.05.. A significant interaction was not observed in an earlier experiment, and the effect in this experiment probably reflects the decreased stimulus-contact time in the quiet CRFko mice, which was not itself statistically significant. Restraint significantly decreased the number of compartment entries
16
A.J. Dunn, A.H. Swiergielr Brain Research 845 (1999) 14–20
tion Ž F1,29 s 0.. The total number of are entries, a useful indicator of locomotor activity was not affected by genotype ŽWT 26.8 " 5.2; CRFko 27.3 " 6.0, F1,29 s 0.3., but was significantly decreased by restraint ŽWT 17.7 " 7.2; CRFko 19.3 " 5.0, F1,29 s 17.4, p - 0.001. and there was no significant restraint= genotype interaction Ž F1,29 s 0.06.. 3.3. BehaÕioral obserÕations in the OF The mice were observed in the OF immediately after completion of an EPM test. ANOVA revealed no significant effects of restraint Žcrossings: F1,36 s 2.7; rears: 1.39. or genotype Ž F1,36 s 1.40 and 1.33. and no significant restraint= genotype interaction Ž F1,36 s 0.58 and 0.26. on the number of crossings and rears in the OF in the first trial. In a second test, the CRFko mice displayed more line crossings than WT mice Ž F1,36 s 4.6, p - 0.05.. There were no other significant differences between the genotypes and treatments. When the data from both trials were pooled ŽFig. 4., CRFko mice displayed more locomotor
Fig. 1. Effect of restraint on the mean stimulus-contact time and the number of compartment entries in the MCC. Wild type and CRF-knockout mice were restrained for 30 min and immediately placed in the MCC and observed for 25 min. UUU Significantly different from unrestrained mice ŽUU p- 0.01, UUU p- 0.001, N s8–9..
Ž F1,30 s 4.36, p - 0.05., but there was no statistically significant difference between the two strains Ž F1,30 s 1.11. and no significant restraint= genotype interaction Ž F1,30 s 1.35.. Also in the MCC, administration of mIL-1b Ž100 ng i.p.. significantly reduced the mean time per contact ŽFig. 2; F1,36 s 50, p - 0.0001.. However, there was no effect of genotype Ž F1,36 s 0.47. and no mIL-1b = genotype interaction Ž F1,36 s 2.1.. In this experiment, none of the other behaviors scored was significantly affected by the genotype or the mIL-1b. 3.2. BehaÕioral obserÕations in the EPM Mice were tested in the EPM the day after a session in the MCC ŽFig. 3.. The number of entries into the open arms was decreased by restraint Ž F1,29 s 14.2, p - 0.001., but there was no effect of genotype Ž F1,29 s 0.3., nor a restraint = genotype interaction Ž F1,29 s 0.. Restraint marginally decreased the time spent on the open arms Ž F1,29 s 4.02, p s 0.054., but there was no significant restraint= genotype interaction Ž F1,29 s 0.02., nor any effect of genotype Ž F1,29 s 0.34.. The number of entries into the enclosed arms was also decreased by restraint Ž F1,29 s 9.0, p - 0.01., but there was no significant effect of genotype Ž F1,29 s 0.0., nor a restraint= genotype interac-
Fig. 2. Effect of mIL-1b on mean stimulus-contact time and the number of compartment entries in the MCC. WT and CRFko mice were injected i.p. with mIL-1b Ž100 ngrmouse. or saline and 90 min later placed in the MCC for 25 min. Significantly different from saline-injected mice ŽUUU p - 0.001, N s10..
A.J. Dunn, A.H. Swiergielr Brain Research 845 (1999) 14–20
17
straint = genotype interactions Ž F1,30 s 0.04 and F1,30 s 1.04.. Grooming was also increased by restraint in the EPM. The total duration increased ŽFig. 5; F1,29 s 7.2, p - 0.05., and the number of bouts was increased Ž F1,29 s 7.1, p - 0.05.. In both cases, there was a significant effect of genotype such that CRFko mice displayed more grooming than WT Ž F1,29 s 11.0, p - 0.01 and F1,29 s 12.9, p - 0.01., but there were no statistically significant restraint = genotype interactions Ž F1,29 s 0.3 and 1.2.. Grooming was not scored in the OF. 3.5. Plasma corticosterone Plasma concentrations of corticosterone were measured following 20 min footshock, and again 2 h after i.p. injection of 100 ng mIL-1b. CRFko mice had substantially lower concentrations of plasma corticosterone compared to WT mice. Footshock elicited the expected increase in WT mice, but no increase was observed in the CRFko mice ŽFig. 6.. ANOVA indicated a statistically significant geno-
Fig. 3. Effect of restraint on behavior in the EPM. WT and CRFko mice were restrained for 30 min and immediately placed in the EPM for 5 min. Top graph: number of open arm entries; Middle graph: time spent on the open arms; Bottom graph: number of enclosed arm entries. U Significantly different from unrestrained mice ŽU p- 0.05, UU p- 0.01, N s8–9..
activity in the OF compared to WT mice, a slightly higher number of line crossings Ž F1,72 s 5.9, p - 0.05. and rears Ž F1,72 s 3.1, p - 0.08., but no genotype= restraint interaction Ž F1,72 s 0.0 and 1.44; crossings and rears, respectively.. 3.4. Grooming behaÕior Restraint increased grooming in the MCC; both the total duration ŽFig. 5; F1,30 s 14.2, p - 0.001. and the frequency Žthe number of grooming bouts; F1,30 s 10.2, p 0.01.. The total duration of grooming was significantly elevated in the CRFko compared to the WT mice Ž F1,30 s 4.52, p - 0.05., but there was no effect of genotype on the frequency Ž F1,30 s 1.25.. There were no significant re-
Fig. 4. Effect of restraint on behavior in the OF. WT and CRFko mice were restrained for 30 min and tested in the EPM for 5 min, after which they were immediately placed in the OF for 5 min. Upper graph: the number of line crossings; lower graph: the number of rears. Data were pooled from two separate experiments. There was a significant effect of genotype Ž p- 0.05. on the number of line crossings Ž N s10..
18
A.J. Dunn, A.H. Swiergielr Brain Research 845 (1999) 14–20
4. Discussion
Fig. 5. Effect of restraint on grooming behavior in the MCC Župper graph. and EPM Žlower graph.. The total time spent grooming observed during the observation periods in the MCC Ž25 min. and the EPM Ž5 min.. Data taken from the experiments in Figs. 1 and 3. U Significantly different from unrestrained mice ŽU p- 0.05, N s8–9..
type = treatment interaction Ž F1,31 s 193, p - 0.0001., reflecting the lack of response in the CRFko mice. IL-1b treatment increased plasma corticosterone in the WT mice, but the response was dramatically reduced in the CRFko mice ŽFig. 6.. ANOVA indicated a statistically significant genotype= treatment interaction Ž F3,28 s 44, p - 0.0001.. The small increase induced by IL-1b in the CRFko mice was statistically significant Ž t 13 s 2.31, p - 0.05..
Behavior in the MCC has been shown to be reliably affected by a variety of stressors, such that the stimuluscontact time is reduced without consistent changes in locomotor activity in rats w1x and mice w3,5x. These responses were closely mimicked by i.c.v. administration of low doses of CRF in mice w3,5x and rats w24x. I.c.v. pretreatment with the CRF-receptor antagonist, a-helical CRF9 – 41 , prevented the restraint-induced changes in behavior w4x, suggesting that brain CRF plays an important role in the stress-related alterations of behavior in the MCC. Furthermore, administration of IL-1 either i.p. w9x or i.c.v. w25x elicited responses in the MCC very similar to those of restraint and i.c.v. CRF. These responses were also attenuated by i.c.v. administration of a-helical CRF9 – 41 w9,25x. The data from the present experiments are similar to those observed earlier in CD-1 mice. Similar behavioral scores were recorded, and there were similar responses to restraint. Mice that lacked the CRF gene, and hence presumably the capacity to synthesize CRF, showed behavioral responses to restraint in the MCC that very closely resembled those of wild type animals. These results confirm those of an earlier experiment in which the behavior in the MCC of CRFko mice was indistinguishable from that of WT w31x. In those experiments, we also observed very similar responses to i.c.v. administration of 20 ng CRF, suggesting that in CRFko mice there is no substantial alteration of the activity of receptors acted upon by i.c.v. CRF. A similar lack of difference between the genotypes was observed in the EPM. Consistent with previous observations w13x, restraint decreased open arm entries and the time spent on the open arms Žalthough the latter effect was only marginally statistically significant.. However, the behavior of the CRFko mice was very similar to that of the WT mice, consistent with results obtained in another labo-
Fig. 6. Effect of footshock and mIL-1b on plasma corticosterone. Blood was collected from WT and CRFko mice Ž n s 17. immediately after removal from their home cages, and then again on the following day immediately after the mice received footshocks for 30 min. Later, the mice were injected with 100 ng mIL-1b i.p. or saline and trunk blood collected following decapitation 2 h later. UUU Significantly different from quiet or saline-injected mice Ž p - 0.0001..
A.J. Dunn, A.H. Swiergielr Brain Research 845 (1999) 14–20
ratory w31x. The total number of entries into both the open and closed arms, a good measure of the overall locomotor activity, was almost identical in both genotypes, with or without restraint. Even though the effects of restraint or genotype might have been affected by the previous exposure of all animals to the EPM, the CRFko and WT mice displayed very similar behavioral response to restraint in the OF. The major distinction between the genotypes was a slightly higher number of line-crossings and rears in the CRFko mice, a response resembling that observed after immobilization, footshock or i.c.v. administration of CRF w16x. Together with the locomotor activity data from the MCC and EPM, these results strongly suggest that the general motor performance is not affected by the CRF deficiency. We observed a significant increase in grooming behavior in CRFko mice relative to WT in both the MCC and in the EPM. Increased grooming in the EPM was also observed in CRFko mice in another laboratory w31x. Grooming can be induced in rats and mice by i.c.v. CRF w11x or ACTH w12x. The grooming associated with the mild stress of novelty may be mediated by ACTH w12x. The enhanced response to CRF might be expected if the lack of CRF in CRFko mice were to result in a supersensitivity of CRF receptors. However, no difference in grooming was observed in the MCC in response to i.c.v. CRF ŽDunn and Swiergiel, unpublished observations.. It is possible that the increased grooming in the CRF-deficient mice is related to compensatory changes in other neuropeptide mediators of grooming, such as oxytocin, prolactin or b-endorphin w7x. The CRFko mice displayed responses to mIL-1b in the MCC that closely resembled those of WT mice, indicating that the ability of IL-1 to induce behavioral responses was not impaired, even though these responses in CD-1 mice appear to depend, at least in part, on functional CRF-receptors w9,25x. We also found that the anorectic response to IL-1 was unimpaired in CRFko mice w27x. Consistent with earlier observations in CRFko mice w18x, we observed low basal concentrations of plasma corticosterone. Footshock elevated plasma corticosterone in the WT mice, but there was no such increase in the CRFko mice. The response to IL-1b in the CRFko mice was substantially reduced, but was still statistically significant. A statistically significant increase of plasma corticosterone following IL-1b was observed in another experiment with CRFko mice, suggesting that this response is real. It is possible that catecholamines or some other peripheral factor was responsible for this elevation. Thus, in contrast to the very small changes observed in behavior, the HPA axis activation in CRFko mice was dramatically impaired. Mice lacking the gene for CRF-1 receptor exhibit decreased stress-related release of ACTH and corticosterone w23,28x. However, they also exhibit decreased anxiety in defensive withdrawal and the EPM w23x and in the light– dark box w28x. In the OF, small increases in locomotor
19
activity were noted w28x. These results are consistent with those reported here in that the stress-related behavioral responses were not absent. Thus, CRF and the CRF-1 receptor appear to be critical for the activation of ACTH and corticosterone secretion, but neither is essential for the behavioral responses. Apparently, other receptors Že.g., CRF-2 receptors. can assume some of the behavioral functions of CRF-1 receptors in the CNS. The general conclusion of these studies must be that CRF-knockout mice display behavior and behavioral responses to stressors remarkably similar to those of wild type animals. This conclusion runs counter to the inference from a large number of studies that brain CRF plays a major role in the behavioral responses normally observed during stress w10,20x. Because attenuation or prevention of the normal response during stress has been observed using a variety of different CRF-receptor antagonists and antibodies to CRF and its receptors, it is unlikely that these results are artifactual. The most obvious explanation is that mice that develop in the absence of CRF acquire other mechanisms to perform some of the normal functions of CRF in wild-type animals. Given that CRF receptors are present in CRFko mice, it is possible that another substance may assume some of the functions of CRF in WT animals. An obvious candidate is urocortin which is known to bind to the known CRF-receptors with affinities similar to those of CRF w30x. However, the anatomic distribution of urocortin does not correlate at all well with that of CRF-receptors w30,31x. Neither urotensin nor sauvagine, both of which bind CRF-receptors, has been detected in mice or rats. Perhaps some undiscovered peptide or nonpeptide ligandŽs. israre involved. Urocortin does not appear to substitute for CRF in the activation of the HPA axis, because, although it is a potent stimulator of ACTH secretion from the pituitary w30x, little, if any, stress-related elevation of plasma ACTH or corticosterone is observed in CRFko mice ŽRef. w18x and Fig. 6.. An alternative possibility is that CRF does not mediate the behavioral responses to stressors in WT mice. Perhaps, urocortin, or some other molecule, performs this function by acting on receptors that can also be activated by CRF. This would explain both the effects of i.c.v. CRF, and the ability of antagonists of presumed CRF receptors to attenuate the responses in stress. The most parsimonious conclusion is that redundancy exists in the CNS mechanisms of stress-related behaviors, although receptors that can be activated by CRF are involved. Tentatively we can conclude that there may be redundancy in both the ligands for CRF receptors, and in the receptors themselves.
Acknowledgements This research was supported by grants from the US National Institute of Mental Health ŽMH46261. and the US National Institute of Neurological Diseases and Stroke
20
A.J. Dunn, A.H. Swiergielr Brain Research 845 (1999) 14–20
ŽNS25370.. We thank Dr. Joseph A. Majzoub and Dr. Louis Muglia ŽChildren’s Hospital, Boston, MA. for providing the homozygous CRF knockout mice Ž crfyry . and the wild type Ž crfqrq . mice. The technical assistance of Michael Adamkiewicz and Galina Mikhaylova is greatly appreciated.
w17x
w18x
w19x
References w20x w1x A.F.T. Arnsten, C. Berridge, D.S. Segal, Stress produces opioid-like effects on investigatory behavior, Pharmacol. Biochem. Behav. 22 Ž1985. 803–809. w2x A.T. Arnsten, D.S. Segal, Naloxone alters locomotion and interaction with environmental stimuli, Life Sci. 25 Ž1979. 1035–1042. w3x C.W. Berridge, A.J. Dunn, Corticotropin-releasing factor elicits naloxone-sensitive stress-like alterations in exploratory behavior in mice, Regul. Pept. 16 Ž1986. 83–93. w4x C.W. Berridge, A.J. Dunn, A corticotropin-releasing factor antagonist reverses the stress-induced changes of exploratory behavior in mice, Horm. Behav. 21 Ž1987. 393–401. w5x C.W. Berridge, A.J. Dunn, Restraint-stress-induced changes in exploratory behavior appear to be mediated by norepinephrinestimulated release of CRF, J. Neurosci. 9 Ž1989. 3513–3521. w6x E.B. De Souza, T.R. Insel, M.H. Perrin, J. Rivier, W.W. Vale, M.J. Kuhar, Corticotropin-releasing factor receptors are widely distributed within the rat central nervous system: an autoradiographic study, J. Neurosci. 5 Ž1985. 3189–3203. w7x F. Drago, A. Contarino, L. Busa, ` The expression of neuropeptide-induced excessive grooming behavior in dopamine D1 and D 2 receptor-deficient mice, Eur. J. Pharmacol. 365 Ž1999. 125–131. w8x A.J. Dunn, Stress-related changes in cerebral catecholamine and indoleamine metabolism: lack of effect of adrenalectomy and corticosterone, J. Neurochem. 51 Ž1988. 406–412. w9x A.J. Dunn, M. Antoon, Y. Chapman, Reduction of exploratory behavior by intraperitoneal injection of interleukin-1 involves brain corticotropin-releasing factor, Brain Res. Bull. 26 Ž1991. 539–542. w10x A.J. Dunn, C.W. Berridge, Physiological and behavioral responses to corticotropin-releasing factor administration: is CRF a mediator of anxiety or stress responses?, Brain Res. Rev. 15 Ž1990. 71–100. w11x A.J. Dunn, C.W. Berridge, Y.I. Lai, T.L. Yachabach, CRF-induced excessive grooming behavior in rats and mice, Peptides 8 Ž1987. 841–844. w12x A.J. Dunn, E.J. Green, R.L. Isaacson, Intracerebral adrenocorticotropic hormone mediates novelty-induced grooming in the rat, Science 203 Ž1979. 281–283. w13x A.L. Gorman, A.J. Dunn, b-Adrenergic receptors are involved in stress-related behavioral changes, Pharmacol. Biochem. Behav. 45 Ž1993. 1–7. w14x S.C. Heinrichs, M.P. Stenzel-Poore, L.H. Gold, E. Battenberg, F.E. Bloom, G.F. Koob, W.W. Vale, E. Merlo Pich, Learning impairment in transgenic mice with central overexpression of corticotropin-releasing factor, Neurosci. 74 Ž1996. 303–311. w15x G.F. Koob, F.E. Bloom, Corticotropin-releasing factor and behavior, Fed. Proc. 44 Ž1985. 259–263. w16x E.H.Y. Lee, Y.P. Tang, C.Y. Chai, Stress and corticotropin-releasing
w21x
w22x
w23x
w24x
w25x
w26x
w27x
w28x
w29x
w30x
w31x
factor potentiate center region activity of mice in an open field, Psychopharmacology 93 Ž1987. 320–324. E.M. Merlo Pich, G.F. Koob, M. Heilig, F. Menzaghi, W. Vale, F. Weiss, Corticotropin-releasing factor release from the mediobasal hypothalamus of the rat as measured by microdialysis, Neuroscience 55 Ž1993. 695–707. L. Muglia, L. Jacobson, P. Dikkes, J.A. Majzoub, Corticotropin-releasing hormone deficiency reveals major fetal but not adult glucocorticoid need, Nature 373 Ž1995. 427–432. L. Muglia, L. Jacobson, J.A. Majzoub, Production of corticotropinreleasing hormone-deficient mice by targeted mutation in embryonic stem cells, Ann. N. Y. Acad. Sci. ŽU.S.A.. 780 Ž1996. 49–59. M.J. Owens, C.B. Nemeroff, Physiology and pharmacology of corticotropin-releasing factor, Pharmacol. Rev. 43 Ž1991. 425–474. E. Potter, S. Sutton, C. Donaldson, R. Chen, M. Perrin, K. Lewis, P.E. Sawchenko, W. Vale, Distribution of corticotropin-releasing factor receptor mRNA expression in the rat brain and pituitary, Proc. Natl. Acad. Sci. U.S.A. 91 Ž1994. 8777–8781. P.E. Sawchenko, L.W. Swanson, Organization of CRF immunoreactive cells and fibers in the rat brain: immunohistochemical studies, in: E.B. De Souza, C.B. Nemeroff ŽEds.., Corticotropin-Releasing Factor: Basic and Clinical Studies of a Neuropeptide, CRC Press, Boca Raton, FL, 1990, pp. 29–52. G.W. Smith, J.-M. Aubry, F. Dellu, A. Contarino, L.M. Bilezikjian, L.H. Gold, R. Chen, Y. Marchuk, C. Hauser, C.A. Bentley, P.E. Sawchenko, G.F. Koob, W. Vale, K.-F. Lee, Corticotropin releasing factor receptor 1-deficient mice display decreased anxiety, impaired stress response, and aberrant neuroendocrine development, Neuron 20 Ž1998. 1093–1102. F. Spadaro, C.W. Berridge, H.A. Baldwin, A.J. Dunn, Corticotropin-releasing factor acts via a third ventricle site to reduce exploratory behavior in rats, Pharmacol. Biochem. Behav. 36 Ž1990. 305–309. F. Spadaro, A.J. Dunn, Intracerebroventricular administration of interleukin-1 to mice alters investigation of stimuli in a novel environment, Brain Behav. Immunol. 4 Ž1990. 308–322. M.P. Stenzel-Poore, S.C. Heinrichs, S. Rivest, G.F. Koob, W.W. Vale, Overproduction of corticotropin-releasing factor in transgenic mice: a genetic model of anxiogenic behavior, J. Neurosci. 14 Ž1994. 2579–2584. A.H. Swiergiel, A.J. Dunn, CRF-deficient mice respond like wild type mice to hypophagic stimuli, Pharmacol. Biochem. Behav. 64 Ž1999. 59–64. P. Timpl, R. Spanagel, I. Sillaber, A. Kresse, J.M.H.M. Reul, G.K. Stalla, V. Blanquet, T. Steckler, F. Holsboer, W. Wurst, Impaired stress response and reduced anxiety in mice lacking a functional corticotropin-releasing hormone receptor 1, Nat. Genet. 19 Ž1998. 162–166. W. Vale, J. Spiess, C. Rivier, J. Rivier, Characterization of a 41-residue ovine hypothalamic peptide that stimulates secretion of corticotropin and b-endorphin, Science 213 Ž1981. 1394–1397. J. Vaughan, C. Donaldson, J. Bittencourt, M.H. Perrin, K. Lewis, S. Sutton, R. Chan, A.V. Turnbull, D. Lovejoy, C. Rivier, P.E. Sawchenko, W. Vale, Urocortin, a mammalian neuropeptide related to fish urotensin I and to corticotropin-releasing factor, Nature 378 Ž1995. 287–292. S.C. Weninger, A.J. Dunn, L.J. Muglia, P. Dikkes, K.A. Miczek, A.H. Swiergiel, C.W. Berridge, J.A. Majzoub, Stress-induced behaviors require the CRH receptor, but not CRH, Proc. Natl. Acad. Sci. U.S.A. 96 Ž1999. 8283–8288.