Does midbrain urocortin 1 matter? A 15-year journey from stress (mal)adaptation to energy metabolism

Abstract

This review summarizes some of the milestones of the research on the biological functions(s) of midbrain urocortin 1 (Ucn1) since its discovery 15 years ago. Detailed characterization of Ucn1 in the midbrain revealed its overall significance in food intake and regulation of homeostatic equilibrium and mood under stress. In addition, we have recently found a conspicuous alteration in midbrain Ucn1 levels in brains of depressed suicide victims. Furthermore, from the results from the genetically modified animals, a picture is emerging where corticotrophin-releasing factor promotes the initial reactions to stress, whereas Ucn1 seems to be crucial for management of the later adaptive phase. In the case of imbalance in action of these principle stress mediators, vulnerability to stress-related brain diseases is enhanced.

Keywords::

• Urocortin 1
• Edinger–Westphal central projecting neurons
• stress
• depression
• energy metabolism
• sex difference

Introduction

Fifteen years ago, the second member of the corticotrophin-releasing factor (CRF) neuropeptide family, urocortin 1 (Ucn1) was discovered (Vaughan et al. 1995). It has a sequence identity of 45% with CRF (Vaughan et al. 1995). Interestingly, Vaughan et al. (1995) localized Ucn1 to a discrete nucleus in the rostroventral midbrain, and to the Edinger–Westphal (EW) nucleus. This has been later confirmed by showing Ucn1 immunoreactive neurons in the EW nucleus in various mammalian as well as non-mammalian species (Bittencourt et al. 1999; Iino et al. 1999; Kozicz et al. 1998, 2002; Ryabinin et al. 2005). The presence of a stress-related neuropeptide in a nucleus classically associated with parasympathetic oculomotor functions (pupil construction, lens accommodation, and trophic functions of the eyes) not only has resulted in the extension of the classical biological significance of EW nucleus but also has become inevitable to revisit the nomenclature of this midbrain area. To distinguish preganglionic cholinergic neurons directly controlling oculomotor functions, neurons that express Ucn1 and do not project to the ciliary ganglion, but rather possess central projections that were renamed as EW central projecting (EWcp) neurons (Kozicz et al. 2011). Despite the ongoing debate on the nomenclature of this brain area, a significant advance has been made in putting together the puzzle on the biological role(s) of EWcp-Ucn1 in the last 15 years. These efforts have brought us closer to answer the questions that we raised at the end of the 1990s. Does midbrain Ucn1 matter? We do believe it does, but to underpin this statement, this review aims to summarize data showing that midbrain Ucn1 is regulated by basic physiological processes, that Ucn1 itself regulates/modulates various physiological processes, and that midbrain Ucn1 may also contribute to the pathophysiology of human diseases.

Does midbrain UCN 1 change under physiological conditions?

Daily rhythms

Period 2 (Per2) is an important clock gene involved in the regulation of the major circadian clock in the mammalian central nervous system, the suprachiasmatic nucleus. In a recent study, we have found that Per2 is expressed in the EWcp (Gaszner et al. 2009). Both the protein and the mRNA expression of Per2 in EWcp neurons showed a diurnal rhythm, with a minimum at light-off, and a maximum at light-on (Gaszner et al. 2009). Interestingly, a majority of Ucn1 neurons co-expressed Per2, suggesting that the Ucn1 is produced in a diurnal fashion. Indeed, we have found a diurnal rhythm in the number of Ucn1-immunopositive neurons (most probably reflecting change in the abundance of Ucn1 peptide in neurons) and in their Ucn1 peptide content, with a minimum at the beginning of the light period and a maximum at light-off, while the UCN1 expression paralleled that of Per2 rhythm (Gaszner et al. 2009). This apparent phase shift in the abundance of Ucn1 peptide and mRNA could represent a diurnal rhythm in the biosynthetic and secretory activity of EWcp-Ucn1 neurons; at light-on, Ucn1 mRNA is produced at its maximum rate, as well as Ucn1 peptide is made, transported, and released from neurons (lowest number of Ucn1 positive neurons), whereas at light-off, Ucn1 mRNA production is at its lowest, and Ucn1 peptide is stored/accumulated in neurons (highest number of Ucn1 positive neurons). These rhythms in the rat EWcp are accompanied by a diurnal rhythm in plasma corticosterone concentration. Diurnal production of Per2 and Ucn1 in the EWcp may be relevant for the physiological role(s) of Ucn1 (see later).

Sex steroids

The presence of estrogen receptor β (ERβ) in the EW nucleus has been previously demonstrated (Mitra et al. 2003). We confirmed this finding and extended it by demonstrating the absence of ERα in EWcp neurons (Derks et al. 2007, 2009). We also found that the majority of EWcp-Ucn1 neurons co-express ERβ in both rat and mouse (Derks et al. 2007, 2009). Furthermore, Ucn1 mRNA in males was nearly 10 times higher than in females in diestrus and 1.6 times higher than in females in proestrus. This clearly indicates a sex-dependent difference in Ucn1 biosynthetic activity. In line with this notion, it has been demonstrated that estrogen decreases the transcriptional activity of the UCN1 promoter via ERβ (Haeger et al. 2006). As an additional support for a sex-dependent expression of EWcp-Ucn1 is the fact that female rats possess significantly higher levels of Ucn1 immunoreactivity than male rats (Fonareva et al. 2009) and that Ucn1 immunoreactivity decreases during the course of pregnancy in rats (Fatima et al. 2007).

Peripheral metabolic signals

In a recent study, using a novel LepRb-IRES-Cre-EYFP-reporter mouse, substantial LepRb expressions were found in EWcp (Scott et al. 2009), and LepRb is expressed in about 50–60% of EWcp-Ucn1 neurons in both rat and mouse (Xu et al. 2009). Furthermore, Ucn1 is also co-expressed with ghrelin receptor mRNA (unpublished observation by the authors). These observations strongly support the view that EWcp-Ucn1 neurons are under the control of peripheral metabolism-related signals and, thus, would play an important role in controlling/modulating food intake and/or energy balance. In support of this notion, we have recently shown that fasting rats for 2 days causes a marked body weight loss, reduces leptin plasma level in both sexes, and results in a 3.3 times upregulation of Ucn1 mRNA in males but not in females (Xu et al. 2009). We have also shown that EWcp-Ucn1 neurons respond to leptin with the activation of JAK2–STAT3 pathway, in which leptin inhibits the electrical activity of EWcp neurons and systemic leptin administration significantly increases Ucn1 content in EWcp (Xu et al. 2010b). Furthermore, electrical lesioning of the EWcp results in inhibition of food intake (Weitemier and Ryabinin 2005) and that high fat diet decreases Ucn1 mRNA in the EWcp too, concomitant with increased plasma leptin level (Legendre et al. 2007). Collectively, these data show that a negative correlation exists between plasma leptin and Ucn1 specifically in the EWcp. Based on the anorexigenic action of centrally administered Ucn1 (Spina et al. 1996), one would expect the opposite. However, one should consider that Ucn1 administered via an intracerebroventricular route has access to both CRF receptor subtypes, which could lead to differential (even opposing) behavioral, physiological, and neuroendocrine responses.

Stress response

Stress is a physiological response to any threatening demand in order to restore homeostatic equilibrium (Selye 1951). Soon after the discovery of Ucn1 in EWcp, it was shown that Ucn1 neurons were recruited by acute pain stress (Kozicz et al. 2001), and that Ucn1 mRNA is significantly upregulated by acute (pain and restraint) stress (Weninger et al. 2000; Kozicz et al. 2001). Later it has become clear that EWcp-Ucn1 neurons are recruited by various acute stressors (Gaszner et al. 2004; Kozicz 2007, 2009), and their activation pattern suggest that they preferentially respond (as revealed by expression of Fos protein) to psychological/processive stressors (e.g. restraint and pain), but not to systemic stressors (hyperosmotic or hemorrhage stress). In addition, in peritoneal inflammation as well as surgical operation (bilateral nephrectomy) both forms of an acute stress, induce strong Fos and Fra-2 expression in EWcp (Lanteri-Minet et al. 1993; Palkovits et al. 2009).

Interestingly, the activation of EWcp neurons and the upregulation of Ucn1 mRNA after acute pain stress last more than 16 h (Kozicz et al. 2001), in contrast to the relative short-lasting (up to 2–4 h) activation of neurons in the hypothalamic paraventricular nucleus (PVN) and the upregulation of PVN-CRF mRNA (Viau and Sawchenko 2002). A similar difference in the dynamics of activation of EWcp and PVN has been found after acute renal failure. EWcp neurons were recruited within 30 min, however, unlike the Fos/Fra-2 expression in the PVN, and EWcp activation lasted at least 72 h during acute renal failure (Palkovits et al. 2009). This not only further supports the view that EWcp neurons are sensitive of an altered body homeostasis but also supports the notion that EWcp is rather involved in the later, adaptive phase of the stress response (Kozicz 2007).

Similarly to rats, restraint stress also increases the expression of egr-1 (another inducible transcription factor) in bird EWcp-Ucn1 neurons (Cunha et al. 2007). In mice, however, no induction of fos or egr-1-ir has been reported following various acute stressors (Turek and Ryabinin 2005; Spangler et al. 2009), suggesting that the stress responsiveness of EWcp neurons is species specific too. In agreement with this idea, electrolytic lesions of this brain area in the mouse, which produce several other robust behavioral effects, do not affect anxiety-like behavior in the plus maze, locomotor behavior in the open field, or levels of plasma corticosterone (Weitemier and Ryabinin 2005).

Chronic stress also recruits Ucn1 neurons in the rodent EWcp. Both chronic predictable/homotypic (chronic ether stress for 21 days in the mouse) and chronic unpredictable/heterotypic (chronic variable mild stress for 14 days in the rat) activated Ucn1 neurons as demonstrated by Fos immunoreactivity (Korosi et al. 2005; Xu et al. 2010a). These data show that EWcp neurons do not habituate to chronic stressors, and this is in clear contrast to the habituating response of PVN neurons upon chronic stress (Viau and Sawchenko 2002).

With respect to Ucn1 peptide and mRNA expression in EWcp neurons, it has to be noted that chronic homotypic stressor induces different dynamics as compared to chronic heterotypic stressor. Chronic ether stress (3 weeks) did not result in any change in Ucn1 mRNA expression in the mouse (Korosi et al. 2005). This response is very similar to that of CRF neurons in PVN, which also did not show any change in CRF mRNA expression (e.g. Romeo et al. 2007). In contrast, chronic unpredictable/heterotypic stress in mouse results in a significant increase in both Ucn1 peptide and mRNA expression (Derks 2010), similarly to the increased CRF mRNA expression after chronic variable stress (e.g. Herman et al. 1995, 2003).

Taken together, the differences in the dynamics of activation of PVN and EWcp neurons both in response to acute and chronic challenges, the hypothesis has been put forward that the complementary action of these stress-sensitive brain areas would serve to terminate the central stress response, thereby promoting successful adaptation to challenges (Weninger et al. 1999; Kozicz 2007, 2009). Consequently, this would imply that besides the CRF neurons in the PVN, EWcp-Ucn1 neurons have an important role in the stress adaptation response and in stress-related brain diseases such as anxiety and depression (Kozicz 2007, 2009).

Does midbrain U_CN1 contribute to disease pathology?

Anxiety and depression

Ucn1, similarly to CRF, administered intracerebroventricularly is capable of accessing its cognate receptors (Bittencourt and Sawchenko 2000) and elicits strong anxiogenic properties (Skelton et al. 2000; Gysling et al. 2004; Kozicz 2007). In addition, a large body of pharmacological evidence indicates specific roles for EWcp-Ucn1 neurons in anxiety-like behaviors. Specifically, the recruitment of EWcp neurons after intracerebroventricular administration of benzodiazepines and selective agonists of the metabotropic glutamate receptors strongly suggests that their action may manifest, at least in part, via regulation of EWcp neuron activity (Linden et al. 2004, 2006; Skelton et al. 2004).

To further dissect the possible role of Ucn1 in disease pathology, with special emphasis on stress-related anxiety and depression, various mice models have been created with a null mutation in the UCN1 gene. One of these models showed significantly heightened anxiety-like behavior, but a normal HPA-axis response to stress, and no change in CRF mRNA expression (Vetter et al. 2002). Another Ucn1-null mice generated by Wang et al. (2002) exhibit normal stress-induced responses, including anxiety-like behavior, autonomic control, and endocrine secretion. These clearly contrasting results make it difficult to conclude about the significance of Ucn1 in regulating anxiety-like behavior. Although some of these contrasting results on Ucn1's role in anxiety-like behaviors can be explained by developmental compensatory mechanisms activated in these mutants (e.g. altered expression in CRF receptors), further and more detailed dissection of the behavioral phenotype of UCN1-null mice and the use of site-specific, lentiviral modification of Ucn1 expression in EWcp are warranted to clarify its biological significance in stress-related behaviors and are the focus of future research in our group. However, it is also possible that in contrast to what was found for mice deficient for receptors CRF-R1 and CRF-R2 (Bale and Vale 2004; Bale 2006), neither CRF nor Ucn1 alone plays an essential role in acute stress-related behavioral and/or autonomic responses. This suggests that CRF and Ucn1 are redundant with respect to their functions in regulating/controlling acute stress response. Instead, Ucn1 may play an important role in regulating the HPA-axis' response to chronic challenges. In line with this notion, another model with a null mutation in the UCN1 gene reveals impaired adaptation to repeated restraint stress (Zalutskaya et al. 2007). Furthermore, in a recent study, Neufeld-Cohen et al. (2010) demonstrated decreased anxiety and increased forebrain serotonergic functioning in UCN1/UCN2 double knockout mice. This together with the fact that Ucn1 neurons project to the dorsal raphe nucleus (DRN; Bittencourt et al. 1999; Weitemier and Ryabinin 2005) clearly puts the Ucn1 neurons in EWcp in critical position to control dorsal raphe serotonergic activity and thereby modulating the stress response and stress-related behavior(s) (Kozicz 2010). Thus, if the lack of innervations by Ucn1/Ucn2 to the DRN results in increased serotonin releases in forebrain anxiety circuitries, it is reasonable to assume that overexpression of UCN1 in EWcp would increase anxiety and depression due to decreased serotonergic activity in the forebrain. And indeed, male depressed suicide victims show an approximately 10 times higher UCN1 expression in EWcp compared to controls (Kozicz et al. 2008).

Although the above data strongly implicate EWcp-Ucn1 DRN–serotonin interaction in controlling/modulating anxiety and depression-like behaviors, the projections of EWcp-Ucn1 neurons to the lateral septum (Bittencourt et al. 1999), which has also been postulated in stress-related anxiety (Koolhaas et al. 1998; Eckart et al. 1999), may also play a specific role.

Mechanistically, dysregulation of the CRF receptor signaling may play a role in stress-related depression. It is well documented that CRF is hypersecreted in forebrain centers and has been suggested to contribute to depression pathology (Holsboer et al. 1984; Nemeroff 1988; Reul and Holsboer 2002; de Kloet et al. 2005; Bale 2006; Joels and Baram 2009). However, current concepts concerning G protein-coupled receptor regulation (Lefkowitz and Shenoy 2005) state that agonist-activated CRF1 receptors undergo rapid desensitization and internalization in response to high agonist concentrations (for review, see Hauger et al. 2006). Therefore, it is unlikely that CRF hypersecretion alone is sufficient to account for the neuropathology of major depression (Hauger et al. 2006). Thus, hypothetically, the concomitant hypersecretion of CRF and Ucn1 could produce exaggerated CRF receptor signaling in brain regions mediating the symptoms of depression (Kozicz 2009).

Gender difference in Ucn1 expression in stress-related mood

Although an increased susceptibility of females to depression has been well documented (Kornstein 1997; Frackiewicz et al. 2000; Kudielka and Kirschbaum 2005), the underlying mechanism(s) is/are not well known. Possibly, gender-specific actions of EWcp-Ucn1 neurons account for this difference. This idea is in line with the finding that chronic variable mild stress does increase Ucn1 mRNA in male but not in female rats and mouse (Derks 2010). The same phenomenon was found in humans, where UCN1 expression appeared to be strongly upregulated (more than 10 times) in male suicide victims with major depression but no such upregulation was observed in female depressed suicides (Kozicz et al. 2008).

Obesity

Peripheral injection of Ucn1 in nanomolar concentration (0.3–3 nmol) produces high and prolonged inhibitory effects on food intake (Asakawa et al. 1999; Wang et al. 2001), an effect that can also be seen in leptin receptor (ob/ob)-deficient mice (Asakawa et al. 2001). The ability of Ucn1 to reduce food intake was interestingly accompanied by a reduced motivation to eat (Kinney et al. 2001), a phenomenon that will need attention in future studies. Intracerebroventricular administration of Ucn1 also potently reduces food intake in food-deprived and free-feeding rats (Spina et al. 1996). This action of Ucn1 is most probably mediated via endogenous Ucn1 in the ventromedial hypothalamic nucleus (VMH), because injection of Ucn1 into the VMH significantly inhibited food intake over 3 h, but no significant effect was observed following injection into other hypothalamic nuclei. This effect of Ucn1 was completely reversed by injecting an antibody against rat Ucn1 into the VMH (Ohata et al. 2000). Interestingly, the VMH, which expresses predominantly CRF-R2 (Van Pett et al. 2000), is densely innervated by Ucn1 fibers (Kozicz et al. 1998; Bittencourt et al. 1999) most probably originating from EWcp. In addition, intracerebroventricular injection of Ucn1 results in strong Fos labeling in VMH (Bittencourt and Sawchenko 2000). Taken together, the function of Ucn1 controlling food intake might be mediated by a EWcp-VMH neuronal circuitry.

To conclude, these data strongly suggest a role for central Ucn1 in regulating food intake and possibly obesity. To underpin this notion, Delplanque et al. (2002) analyzed the association between single nucleotide polymorphisms in the UCN1 gene and obesity in French Caucasians. No strong associations have been found (at least not in French Caucasians), suggesting that the UCN1 gene has a major role in obesity (Delplanque et al. 2002). This clearly contrasts the evidence of linkage of obesity-related phenotypes (i.e. leptin resistance, body mass index) to human chromosome 2 at a position of 2p21–23, a locus where two potentially interesting genes lie, i.e. UCN1 and proopiomelanocortin (Comuzzie et al. 1997; Rotimi et al. 1999; Delplanque et al. 2002). It is, however, possible that Ucn1 indirectly affects the development of obesity. Synergistic effects of leptin and Ucn1 on food intake have been shown in several studies. More specifically, Kotz et al. (2002) showed that the satiety effect of Ucn1 is enhanced by its induction of plasma leptin. Also, co-treatment of rats with doses of leptin and Ucn1 that are ineffective when given alone effectively suppresses appetite (Pan and Kastin 2008). Furthermore, leptin shows a facilitatory effect on Ucn1 transport across the blood–brain barrier, and there is evidence that Ucn1 potentiates leptin signaling by increasing leptin receptor-induced STAT3 phosphorylation (Pan et al. 2007). Such an action of Ucn1 may amplify the cellular response of leptin and suggests that Ucn1 can play an important compensatory role during leptin resistance in obesity (Pan and Kastin 2008).

Beside Ucn1's involvement in regulating food intake, Ucn1 has also been implicated in another consumptive behavior, i.e. intake of alcohol and addictive drugs. This interesting aspect of the biology of midbrain Ucn1 is, however, beyond the scope of this review and has recently been reviewed (e.g. Ryabinin and Weitemier 2006; Vilpoux et al. 2009; Ciccocioppo et al. 2009; Kozicz et al. 2010; Lowery and Thiele 2010). Collectively, data on addictive drugs and eating behavior suggest that EWcp-Ucn1 could have a unique role in consumption of rewarding substances.

A model

On the basis of the data mentioned above, we have constructed a model of the possible biological significance of midbrain Ucn1 neurons in the EWcp (Figure 1). These neurons receive multiple inputs transmitting various sorts of information ranging from endocrine (e.g. estrogen, cortisol), metabolic (e.g. leptin, ghrelin), nociceptive, and visceral signals (Morimoto et al. 1996; Derks et al. 2007, 2009; Gaszner et al. 2007; Xu et al. 2009). These will be received by receptors for estrogen (ERβ; Derks et al. 2009), cortisol (GR; data under publication), ghrelin (GHSR; Kaur and Ryabinin 2010), leptin LepRB (Xu et al. 2009), and neuropeptide Y (NPY-Y1 and Y5; Gaszner et al. 2007). EWcp-Ucn1 neurons integrate these diverse input signals, which will consequently alter their secretory activity. This will result in changes in the synaptic release of Ucn1 in brain areas targeted by Ucn1 afferents, including the lateral septum (LS), DRN, and VMH (Bittencourt et al. 1999; Weitemier and Ryabinin 2005). All these brain nuclei express CRF-R2, a receptor to which Ucn1 binds with a very high affinity (Vaughan et al. 1995), and thus Ucn1 released in these brain areas is in state to influence the functioning of these brain centers. Clearly, these centers have also been implicated in regulating the (patho)physiology of the stress response and food intake, as well as in controlling mood. According to our model, dysfunctioning of EWcp neurons would contribute to dysregulation of neuronal activity in LS, DRN, and VMH. This consequently will result in alterations in homeostatic equilibrium and in maladaptive regulation of stress and food intake.

Figure 1.  The proposed model. CRF-R2, corticotrophin-releasing factor receptor type 2; DRN, dorsal raphe nucleus; GHSR, ghrelin receptor; GR, glucocorticoid receptor; ERβ, estrogen receptor β; LS, lateral septal nucleus; NPY-Y1 and Y5, neuropeptide Y receptor Y1 and Y5; ObR, leptin receptor; VMH, ventromedial hypothalamic nucleus.

Does midbrain UCN1 matter?

We do believe it does! However, there are still several pieces of the puzzle that are missing, despite the large body of (mostly correlative) evidence that has been collected in the past 15 years. In the future, we will need comprehensive neuroanatomical studies mapping all afferent and efferent connections to and from EWcp-Ucn1 neurons, identifying the phenotype of neurons targeted by Ucn1 efferents. In addition, further and detailed physiological and behavioral characterization of existing Ucn1-null mice will be inevitable under control and challenge conditions (including yet untested hypotheses, such as reward processing and cognition) to gain better insights into the (patho)physiological significance of Ucn1. Ultimately, viral-mediated down- or upregulation of UCN1, specifically in the EWcp would serve us with direct evidence on the roles of EWcp neurons, so we could confidently state that MIDBRAIN UCN1 DOES MATTER!

Acknowledgements

We apologize to those colleagues whose work could not be cited here owing to space limitations. L.S. and X.L are on a 4-year PhD program sponsored by MSD, Oss, The Netherlands (L.S.), and by Radboud University Nijmegen, Faculty of Science, Mathematics and Informatics (X.L.). This work was also supported by the Netherlands Organization for Scientific Research Grant 864.05.008 (to TK).

Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.