The dysphoric component of stress is encoded by activation of the dynorphin kappa-opioid system

Abstract

Stress is a complex human experience having both positive and negative motivational properties. When chronic and uncontrollable, the adverse effects of stress on human health are considerable and yet poorly understood. Here, we report that the dysphoric properties of chronic stress are encoded by the endogenous opioid peptide dynorphin acting on specific stress-related neuronal circuits. Using different forms of stress presumed to evoke dysphoria in mice, we found that repeated forced swim and inescapable footshock both produced aversive behaviors that were blocked by a kappa-opioid receptor (KOR) antagonist and absent in mice lacking dynorphin. Injection of corticotropin-releasing factor (CRF) or urocortin III, key mediators of the stress response, produced place aversion that was also blocked by dynorphin gene deletion or KOR antagonism. CRF-induced place aversion was blocked by the CRF2 receptor antagonist antisauvigine-30, but not by the CRF1 receptor antagonist antalarmin. In contrast, place aversion induced by the KOR agonist U50,488 was not blocked by antisauvigine-30. These results suggest that the aversive effects of stress were mediated by CRF2 receptor stimulation of dynorphin release and subsequent KOR activation. Using a phospho-selective antibody directed against the activated KOR to image sites of dynorphin action in the brain, we found that stress and CRF each caused dynorphin-dependent KOR activation in the basolateral amygdala, nucleus accumbens, dorsal raphe, and hippocampus. The convergence of stress-induced aversive inputs on the dynorphin system was unexpected, implicates dynorphin as a key mediator of dysphoria, and emphasizes kappa-receptor antagonists as promising therapeutics.

Figure 1.

Stress-induced odorant aversion is dynorphin/KOR mediated. A, Schematic of the odorant-stress aversion paradigm. B, Odorant aversion scores (odorant side minus control side) show that, in the absence of stress pairing, the odorant does not elicit aversion (the odorant is neutral). Mice previously experiencing stress paired with odorant showed significant aversion to subsequent odorant exposure, and norBNI pretreated mice did not develop odorant aversion despite stress pairing. **p < 0.01, significantly different from saline/stress group. C, Pdyn(+/+), but not Pdyn(−/−), mice developed significant aversion to odorant. *p < 0.05, significantly different from Pdyn(−/−) group. D, Pairing of the neutral odorant with cocaine injection produced significant place preference for the odorant compartment that was not blocked by norBNI pretreatment. E, κ-Opioid receptor stimulation caused a dose-dependent increase in the odorant-aversion score. When odorant was paired with either 0, 2.5, 5, or 10 mg/kg of the selective κ-opioid receptor agonist U50,488, mice showed an increased aversion to the odorant when subsequently presented with the odorant alone in the T-maze (n = 4–8 animals per group).

Figure 2.

Footshock stress-induced aversion is mediated by the dynorphin/KOR system. Significant aversion was evident to a compartment previously paired with shock compared with unshocked controls. Unshocked controls spent nearly equal time in both compartments during the 15 min session. Wild-type (WT) mice pretreated with the κ-opioid antagonist norBNI (10 mg/kg, i.p.) did not show significant place aversion after shock training. Mice lacking dynorphin (Pdyn, −/−) also failed to develop a significant place aversion after shock training. **p < 0.01, significantly different from saline/no shock group (n = 5–8 animals per group).

Figure 3.

CRF injection increases KOR phosphorylation. A, Representative Western blots showing that CRF (1 μg/5 μl, i.c.v.) increased the 65 kDa KORp-ir band (lane 2) in striatal extracts compared with saline injection (lane 1). Pretreatment with norBNI (10 mg/kg, i.p.) blocked the CRF-induced increase (lane 3). The bar graph quantifies the increase in KORp-ir intensity and norBNI antagonism evident in replicate gels. *p < 0.05, significantly different from saline (n = 4–12 animals per group). B, Representative images showing a substantial increase in KORp-ir (green) in the basolateral amygdala (BLA) after injection of CRF compared with saline-injected mice. The increased KORp-ir was not evident in mice pretreated with norBNI (CRF + NorBNI) or in Pdyn(−/−) mice injected with CRF. GAD67-ir is also shown (red), and the bottom row of images is shown at higher power. Data represent three independent experiments taken from separate animals.

Figure 4.

The dynorphin/κ-opioid system mediates the aversive properties of CRF. A, Schematic showing the conditioning paradigm used to assess aversion. B, CPA scores (drug compartment posttest minus pretest) showed that mice injected with saline in both compartments did not develop aversion (gray bar, left) and that norBNI (10 mg/kg, i.p.; white bars) with saline did not affect place preference. CRF (1 μg, i.c.v.) injection induced significant CPA (gray bar, right) that was blocked by norBNI pretreatment (white bar). *p < 0.05, significantly different from other groups at (n = 9–16 animals per group). C, CRF (1 μg, i.c.v.) produced significant CPA in wild-type Pdyn (+/+) mice but not in Pdyn(−/−) littermates. *p < 0.05, significantly different from Pdyn(−/−); n = 5–8 animals per group.

Figure 5.

CRF₂-R activation causes dynorphin/KOR-mediated aversive behavioral responses. A, Representative Western blots showing that urocortin III (U3) injection (0.5 μg, i.c.v.) increased KORp-ir compared with saline-injected mice. The increased KORp-ir was blocked by pretreatment with norBNI (10 mg/kg, i.p.) (U3+N) (n = 4 animals per group). B, Urocortin III (UcnIII) induced significant place aversion that was blocked by norBNI, whereas Stressin 1 (0.5 μg, i.c.v.) did not induce significant aversion. *p < 0.05, significantly different from UcnIII group (n = 6 animals per group).

Figure 6.

CRF2-R antagonism blocks CRF-induced, but not U50,488-induced, aversion. A, CRF-induced place aversion was blocked by ASV-30 (1 nmol, i.c.v.) pretreatment, whereas antalarmin (10 mg/kg, i.p.) pretreatment did not. U50,488 also induced a place aversion, and ASV-30 pretreatment did not block this aversion. Veh, Vehicle (see Materials and Methods). *p < 0.05, significantly different from ASV-30 (n = 10–11 animals per group). B, Representative images showing that the increase in KORp-ir (green) in the basolateral amygdala caused by CRF injection (middle panel) compared with saline (left panel) was blocked by pretreatment with ASV-30 (right panel). Counterstained for GAD67-ir (red). C, Injection of the CRF₁-R-selective antagonist antalarmin (10 mg/kg, i.p.) before swim-stress testing significantly blocked stress-induced colonic motility (fecal output) (n = 8; ***p < 0.001; t test) as reported previously (Griebel et al., 2002). D, Injection with antalarmin (10 mg/kg, i.p.) significantly blocked swim stress-induced immobility (n = 4; *p < 0.05; t test), demonstrating that this dose of antalarmin was pharmacologically active.