We created a transgenic model of chronic CRF hypersecretion (CRF-Tg mice) using a chimeric CRF transgene comprised of the metallothionein promoter linked to the rat CRF genomic gene (Stenzel-Poore et al., 1992). These mice exhibit chronic HPA axis activation due to central overproduction of CRF. Basal ACTH is elevated two- to five-fold, resulting in a tenfold elevation in circulating glucocorticoids. Physically, such high corticosterone levels render these mice Cushingoid. By eight weeks of age, CRF-Tg mice present with thin skin, hair loss, brittle bones, truncal obesity, and a characteristic buffalo hump; a phenotype, which can be reversed with adrenalectomy (Stenzel-Poore et al., 1992). CRF is overproduced in CRF-Tg mice in regions of the brain that normally express CRF (e.g., PVN, preoptic area, amygdala, olfactory bulb, and lateral septum). CRF mRNA expression is increased in these regions except the PVN where CRF expression matches control levels. This may reflect downregulation of endogenous CRF in the PVN due to negative feedback via increased circulating glucocorticoids. We anticipated widespread expression of the transgene in the brain and periphery, since the CRF promoter is replaced by the more broadly expressed metallothionein promoter. However, this did not occur, suggesting that sequences located within the rat genomic CRF gene contain information that regulates expression in certain CNS sites and peripheral tissues. In keeping with this, peripheral expression of the mMT-CRF transgene follows a pattern similar to that known for endogenous peripheral CRF localization: lung, adrenal, heart and testis. Notably, circulating CRF is not elevated consistent with the lack of significant CRF production in peripheral sites. The restricted expression profile of the CRF transgene results in a model that mimics the physiology of chronic activation of the HPA axis and CRF dysregulation in the brain.
With such marked increases in basal circulating glucocorticoids, it was of interest to determine whether CRF-Tg mice could further activate the HPA axis in response to acute stress. We find that ACTH is not significantly elevated following restraint stress, in contrast to a tenfold increase in WT littermates (S. Murray and M. Stenzel-Poore, unpublished). Interestingly, CRF-Tg mice also lacked a corticosterone response immediately following restraint stress, however a twofold increase in corticosterone was observed 20min later. Therefore, pituitary-adrenal responses to stress are both protracted and suppressed in CRF-Tg mice. Similarly, following immune stress (i.p. injection of LPS), CRF-Tg mice do not mount a detectable ACTH response, but again show a delayed increase in corticosterone compared with WT mice. These data suggest that chronic HPA activation desensitizes the system to further stimulation by an exogenous stressor. Such reduced pituitary responsiveness in CRF-Tg mice may be a consequence of excess glucocorticoid negative feedback and/or reduced pools of stored ACTH. It is unclear how CRF-Tg mice mount a corticosterone response in the absence of further ACTH induction. Corticosterone elevation may reflect stimulation by subdetectable increases in ACTH or may result from CRF or other stress-reactive mediators (such as IL-6) acting directly on the adrenal.
CRF is widely considered to be a critical mediator of stress-related behaviors via its actions as a neurotransmitter in hypothalamic and extrahypothalamic sites (Dunn and Berridge, 1990). CRF-Tg mice are a unique model to study the effects of excess CRF
on behavior. Most notably, CRF-Tg mice exhibit hypoactivity in novel environments and display increased anxiety-like behavior in paradigms that measure innate anxiety, including the elevated plus maze, light-dark paradigm and novel open field (Stenzel-Poore et al, 1994; Van Gaalen et al, 2002). Importantly, enhanced anxiety is reversed with central administration of CRF receptor antagonists (Stenzel-Poore, 1992) and occurs independent of pituitary-adrenal activation (Heinrichs et al, 1997). Heightened anxiety in these animals appears to be due in part to altered serotonergic function, as cilopram, a selective serotonin reuptake inhibitor, reduces anxiety-like behavior in CRF-Tg mice (M. Van Gaalen, unpublished). In addition, CRF-Tg mice show marked deficits in learning and memory processing that are reversed by pretreatment with a benzodiazepine anxiolytic drug, which suggests that excessive anxiety and arousal in these mice interferes with cognitive function (Heinrichs et al, 1996). It should be noted that not all aspects of anxiety are altered in this model. Anxiety measured in conflict and conditioned fear paradigms were similar between CRF-Tg and WT mice (Van Gaalen et al, 2002, 2003). Thus different components of anxiety may be modulated by different neural pathways.
Both Cushing's and corticosteroid-treated patients frequently experience greater susceptibility to infection, presumably due to glucocorticoid-mediated immunosuppression (Nelson, 1989). We have used CRF-Tg mice to examine the specific effects of moderate but unrelenting elevations in glucocorticoids on immune function. CRF-Tg mice have reduced numbers of leukocytes in the bone marrow, thymus, spleen, and blood (Stenzel-Poore et al, 1996; Murray et al, 2001), which is primarily due to decrease in the number of T and B lymphocytes. In addition, CRF-Tg mice develop lower antibody titers following immunization. This is consistent with the dearth of B lymphocytes; however, CRF-Tg mice display qualitative as well as quantitative deficits in antibody responses following immunization. Isotype switching from IgM to IgG is poor and antibody specificity is altered. Maturation of the antibody response depends on specialized microenvironments within lymphoid tissues, referred to as germinal centers. CRF-Tg mice fail to form germinal centers following immunization. Thus, CRF overexpression alters the ability of the adaptive immune system to respond normally to immune challenge, possibly by altering the microenvironment required to activate antigen-specific lymphocytes. We believe these changes are mediated by increased HPA activation in CRF-Tg mice, because treatment of WT mice with chronic corticosterone has a similar effect.
CRF-overexpressing mice, line 2122 (CRF-OE2122)
Recently, Groenink and colleagues created another model of CRF overexpression, CRF-OE2122 mice (Groenink et al, 2002). In this model, the transgene consists of the coding sequence of rat CRF cDNA linked to the murine Thy-1.2 gene. Regulatory sequences of Thy-1.2 target transgene expression to postnatal and adult neurons, thus CRF overproduction is limited to the CNS (Dirks et al, 2002). Similar to CRF-Tg mice, CRF-OE2122 mice present with heightened HPA axis activity. However, plasma ACTH is normal and basal circulating corticosterone levels are elevated only ~ fourfold (Groenink et al, 2002). Thus, these mice develop mild Cushingoid features (hair loss, fat deposition) only later in life (~6 months). Differences between the two CRF transgenic models are likely due to different promoters, copy numbers, and/or regulatory sequences surrounding the insertion sites.
CRF-OE2122 mice show a similar neuroendocrine profile to that of individuals with major depression. Increased CRF production in the CNS has been implicated in major depression and, like CRF-OE2122 mice, most patients exhibit elevated circulating Cortisol levels in the face of normal ACTH levels (Gold et al, 1986; Chrousos, 1998). In addition, CRF-OE2122 mice show a flattened diurnal rhythm of glucocorticoid secretion and fail to suppress corticosterone secretion in a dexamethasone suppression test (Carroll, 1982; Groenink et al, 2002). These findings suggest that negative feedback of the HPA axis is altered, possibly due to reduced glucocorticoid receptor sensitivity, though this has not been tested formally in these mice. CRF-OE2122 mice do not show the behavioral profile resembling anxiety and/ or depressive illness. These mice respond normally in a number of tasks that measure anxiety- and depressive-like behavior, despite overexpression of CRF in limbic brain regions (Dirks et al., 2001, 2002). However, reduced startle reactivity and prepulse inhibition was recently reported (Dirks et al., 2002). The lack of a marked anxiety-like phenotype is at odds with the expression of such behaviors in CRF-Tg mice or following exogenous CRF administration. Physiologically, these mice show elevated heart rate and body temperature during the light phase of the diurnal cycle, which the investigators suggest may be associated with the need for increased energy intake as CRF-OE2122 show increased food and water consumption without a concomitant increase in body weight gain (Dirks et al., 2002)
Was this article helpful?