Partial knock-down and tissue-specific mutations of the glucocorticoid receptor have circumvented perinatal mortality and allowed for examination of glucocorticoid pathways in adult physiology. Pepin et al. (1992) created a transgenic mouse model that harbors a transgene that constitutively expresses antisense RNA against GR. To restrict transgene expression to the CNS, a neurofilament promoter was used. Thus, antisense expression impairs production of GR mRNA predominantly in neural tissue, however, ectopic expression occurs in the pituitary and several peripheral tissues. GR signaling is only partially impaired in transgenic mice as GR mRNA levels are reduced 50-70% in hypothalamus and only 30-55% in peripheral organs (Pepin et al, 1992), thus, it should be cautioned that changes seen here may be unique. At the onset, it appeared that this model may closely resemble clinical depression in humans, in terms of neuroendocrine function. Transgenic mice display decreased negative feedback efficacy to both corti-costerone and dexamethasone; a tenfold higher dosage of dexamethasone is required to suppress plasma ACTH and corticosterone levels (Stec et al, 1994; Barden et al, 1997). This resistance to the suppressive effects of dexamethasone is similar to human depression, where 60-70% of severe clinical cases are nonsuppressors (Carroll, 1982). Using in vitro methods with hypothalamic-pituitary complexes from transgenic mice, Karanth et al. (1997) have subsequently shown that impaired glucocorticoid negative feedback occurs primarily at the level of the hypothalamus and not the pituitary. In addition, these mice display exaggerated ACTH responses to stress and exogenously administered CRF (Barden et al, 1997; Karanth et al, 1997), whereas corticosterone responses are reduced due to hyposensitivity of the adrenal gland (Barden et al, 1997). However, unlike human depression, which typically presents with elevated urinary free Cortisol levels, and elevated levels of Cortisol and CRF in cerebrospinal fluid, transgenic mice show normal or elevated corticosterone levels only in the morning (Pepin et al, 1992; Karanth et al, 1997). Furthermore, recent studies indicate that transgenic mice show reduced hypothalamic CRF activity rather than CRF overactivity (Dijkstra et al, 1998). Thus, despite some similarities with human depression, these mice do not mimic all neuroendocrine features of the illness.
Behavioral studies revealed spatial learning deficits in these mice (Rousse et al, 1997; Steckler et al, 1999), which supports the notion that GR signaling is involved in specific learning and memory processing, presumably at the level of the hippocampus.
Indeed, hippocampal deficits in long-term potentiation have been observed in transgenic mice (Steckler et al, 2001). In addition, these mice show reduced anxiety-like behaviors based on elevated plus maze testing and behavioral responses recorded during predator stress (Montkowski et al, 1995; Linthorst et al, 2000). It has been suggested that hypo-activity of hypothalamic CRF may account for anxiolytic behavior in this model. Recently, this model has been used extensively to examine the interaction of HPA axis activity with other neurotransmitter systems, particularly the serotonin system. Interestingly, microdialysis studies showed an exaggerated elevation in hippocampal serotonin levels during stress (Linthorst et al, 2000), which suggests that GR signaling may attenuate serotonin responsiveness. In addition, alterations in serotonergic receptor binding were found, though only in hippocampal regions that contain both GR and MR (Farisse et al, 2000). Thus, regulation of serotonin signaling may require both MR and GR.
Brain-specific glucocorticoid receptor knock-out mice (GRNesCre)
Recently, Tranche et al. (1999) generated a conditional GR knockout mouse (GRNesCre mice) where GR function is selectively inactivated in the CNS, using a Cre/loxP-recombination system in which Cre is under the control of the nestin promoter/enhancer. In this model, GR protein is absent in the brain but is normally distributed in the anterior pituitary and other peripheral tissues. GRNesCrc mice display pronounced alterations in HPA axis equilibrium despite intact negative feedback on pituitary cells. Hypothalamic CRF expression is elevated, leading to increased POMC transcription in the pituitary and elevated plasma corticosterone. These finding demonstrate that intact negative feedback at the pituitary is not able to overcome uninhibited hypothalamic CRF drive, because glucocorticoids are still able to act on nonneuronal cells with intact GR, these mice display several symptoms of Cushing's syndrome, including growth retardation, altered fat distribution and osteoporosis. A detailed study of energy regulation in weanling and adult GRNcsCre mice highlights the catabolic signaling of increased hypothalamic CRF and circulating corticosterone (Kellendonk et al, 2002). Despite basal elevations in the HPA system, GRNesCre mice show normal elevations in ACTH and corticosterone following restraint stress. These mice exhibit less anxious behavior in the dark-light box and elevated zero maze paradigms (Tronche et al, 1999), similar to GR antisense transgenic mice. As mentioned above, glucocorticoids have been shown to increase CRF levels in the amygdala, possibly contributing to anxiety-like behavior. Such changes may not occur in these mice lacking GR.
Glucocorticoid receptor DNA binding domain knock-out mice (GRdim)
Tronche et al. also developed a mouse model with a point mutation within the D loop of the second zinc finger of the receptor (Reichardt et al, 1998). This mutation causes loss of dimerization and binding of the receptor to DNA targets, thereby eliminating one mechanism that GR utilize to alter transcription rates. Owing to the precise mutation, modulation of transcription by protein-protein interactions remain intact. Thus, this model has the potential to relate molecular signaling to overt glucocorticoid actions. Interestingly, GRdlm mice show normal viability and lung development, indicating that proteinprotein interactions alone may be sufficient for lung maturation. However, HPA axis regulation and feedback inhibition by glucocorticoids requires DNA binding of GR specifically at the pituitary but not the hypothalamus. Newborn GRdlm mice display normal CRF immunoreactivity in the median eminence; however, POMC mRNA, and ACTH immunostaining were greatly elevated in the anterior pituitary. These findings demonstrate that glucocorticoids utilize different molecular mechanisms to regulate a single physiological system. Interestingly, impaired inhibitory feedback at the pituitary leads to increased and protracted corticosterone levels in response to stress. These mice display normal locomotor activity, exploration and anxiety-related behavior but show impaired performance in a spatial memory task (Oitzl et al, 2001). When considered with the reduced anxiety observed in GRNesCre mice, these findings suggest that emotional versus learning behaviors may be modified by GR activation via different nuclear mechanisms.
A mouse model (YGR mice) of GR overexpression was recently created (Reichardt et al, 2000) by introducing two additional copies of the GR gene into the genome using yeast artificial chromosomes (YAC). The entire gene including regulatory elements responsible for gene transcription and large portions of flanking sequences is contained on a YAC, thus transgenic expression should mimic endogenous GR gene expression. Owing to autoregulatory mechanisms, GR mRNA expression did not reach the theoretical twofold elevation as predicted to occur with four alleles. Thus, GR mRNA was elevated by 60% and 43% in brain and pituitary, respectively, while 20-24% elevation was observed in peripheral tissues of the spleen, thymus, and liver. Importantly, it was shown in hippocampus that these increases in transcription resulted in elevated (~ 50%) protein expression.
As predicted, overexpression of GR resulted in a strong suppression of HPA axis activity, showing the opposite dysregulatory effects of GR knock-out models. Thus, YGR mice show a two- to threefold reduction of immunoreactive CRF content in the median eminence and POMC mRNA expression as well as diminished circulating corticosterone. In response to stress, elevations in corticosterone are smaller and decline more rapidly in YGR mice compared to controls. These results are indicative of enhanced negative regulatory feedback by GR overexpression. Interestingly, these mice illustrate the significant adaptive value of glucocorticoids in suppressing immune and inflammatory responses. YGR mice show increased T cell apoptosis, diminished release of the inflammatory cytokine, IL-6 in response to LPS injection, and increased resistance to endotoxic shock (Reichardt et al, 2000).
We have described various genetic models in which components of the corticotropin system have been individually compromised. This unqiue dissection highlights the role of corticotropin pathways in promoting stability through adaptive changes. In each of these models, allostatic processes are compromised. Certain models show overt changes in basal function as a consequence of HPA axis impairment. In such instances, compensatory mechanisms do not over-ride the deficiency resulting in allostatic load and subsequent pathophysiology. The CRF transgenic mouse reflects a failure to turn off allostatic processes. Thus, adaptive hormones such as glucocorticoids are unable to negatively regulate unmitigated CRF overexpression leading to an elevation in hormones at all levels of the HPA axis. Such allostatic load is readily apparent in the Cushing's phenotype of these mice. In other models, compensatory mechanisms are more effective in maintaining normal function and allostatic load is not evident in the basal state. However, altered allostatic processes become apparent when systems are challenged. These models have contributed greatly to our understanding of the plasticity of neurohormonal regulation. For example, both CRF and CRF| receptor KO mice are unable to elevate glucocorticoids in response to behavioral stress. Yet, robust HPA axis activation is seen following immune challenge, showing that the immune system can directly signal pituitary corticotropes, circumventing hypothalamic input. In addition, these models have shown that different mechanisms of molecular signaling (e.g., DNA binding vs. protein-protein interaction of GR) in part, confer plasticity of HPA axis function. Likewise, different routes of signaling (CRF| receptors vs. CRF2 receptors) appear to provide temporal tailoring of the response to stress. Using these models and others which are sure to follow, we can anticipate increasing clarification of the complexities of the HPA axis and its essential role in maintaining allostasis.
Was this article helpful?