Steroid Hormone Action in the Brain: Cross-Talk Between Signalling Pathways

Abstract

Ovarian steroid hormones, oestradiol and progesterone, modulate neuroendocrine functions in the central nervous system, resulting in alterations in physiology and behaviour. The classical model of steroid hormone action assumes that these neural effects are predominantly mediated via their intracellular receptors functioning as ‘ligand-dependent’ transcription factors in the steroid-sensitive neurones regulating genes and genomic networks with profound behavioural consequences. Studies from our laboratory demonstrate that, in addition to their cognate ligands, intracellular steroid receptors can be activated in a ‘ligand-independent’ manner by the neurotransmitter dopamine, which alters the dynamic equilibrium between neuronal phosphatases and kinases. A high degree of cross-talk between membrane-initiated signalling pathways and the classical intracellular signalling pathways mediates hormone-dependent behaviour in mammals. The molecular mechanisms, by which a multitude of signals converge with steroid receptors to delineate a genomic level of cross-talk in brain and behaviour are discussed.

Ovarian steroid hormones, oestradiol (E₂) and progesterone, play an important role in regulating physiology and reproduction in a wide variety of species (1, 2). One of the well-known ‘classical’ mechanisms by which steroid hormones mediate their biological effects is by binding to their specific intracellular receptors that act as ligand (hormone)-inducible transcription factors, inducing structural and functional changes to facilitate target gene expression and regulation (3–6). However, several studies indicate that not all the biological effects are mediated by direct receptor modulation and target gene expression transcription (7–10). Indeed, there are numerous reports of their rapid ‘nonclassical’ effects on cell membrane/cytoplasmic signal transduction pathways, which are independent of protein synthesis and transcriptional regulation (11–18). In addition to these, several laboratories, including our own, have described a ‘ligand-independent’ mechanism of activation, a phenomenon characterised by the activation of the classic intracellular receptors by factors other than their cognate ligands (19–24). Studies in recent years suggest that these mechanisms are not mutually exclusive, but interact with each other to achieve a physiological end-point. In this review, we discuss our studies on ligand-independent mechanisms of steroid hormone receptor activation with specific reference to the effects of progesterone in brain and behaviour.

Progestin receptor (PR) activation in brain and reproductive behaviour

As in other reproductive tissues, progesterone regulates cellular functions in the central nervous system resulting in alterations in reproductive physiology and behaviour (2, 25). PRs undergo significant conformational change upon binding by progesterone leading to their nuclear translocation, dimerisation and DNA binding (3–5). When bound to DNA, PRs interact with basal transcriptional machinery, assisted by coactivator molecules to initiate chromatin remodelling (26–28). Phosphorylation of the coactivators is also considered to play a crucial role in the activation of PRs (29–31). The time course of activation and termination of sexual behaviour parallels the E₂-induced increase and decline in PRs in the ventrolateral region of the ventromedial hypothalamus and the preoptic area of the brain (32–36). Studies using PR antagonists, protein and RNA synthesis inhibitors, antisense oligonucleotides to PR, and mutant mice with targeted null mutation of the PR gene are consistent with the involvement of PR-mediated classical mechanism of the action of progesterone in mediating sexual behaviour (37–44).

Dopamine activation of PRs in female reproductive behaviour: a ligand-independent mechanism

Although the conventional model of PR activation assumes that progesterone is absolutely required for the activation of the receptor (ligand-dependent), studies in the past decade have shown that, under some circumstances, PRs can be activated by factors other than its cognate ligand, progesterone (19–21). These studies are also substantiated by the in vitro observations that, in the absence of progesterone, 8-bromo cAMP, can mimic the progesterone-dependent, PR-mediated transcription (45). Pharmacological stimulation of cAMP-dependent protein kinase (PKA) mimicked progesterone-dependent, PR-mediated transcription in the absence of progesterone; this effect could be blocked by inhibition of PKA, suggesting that the phosphorylation of the PR or other proteins in the transcription complex can also modulate PR-mediated transcription (45).

Ligand-independent activation of PRs by dopamine (DA) and DA agonists resulting in the translocation of PRs from the cytoplasm to the nucleus was demonstrated in an in vitro cell transfection system (19, 20). In addition, okadaic acid, an inhibitor of protein phosphatases 1 and 2 (PP1 and PP2A) stimulated not only PR-, but also DA-mediated transcription in PR-negative CV1 cells co-transfected with PR expression vector and a reporter plasmid containing a progesterone response element upstream of chloramphenicol acetyltransferase gene in vitro (20). Using female reproductive behaviour of rats and mice as a model, we examined the physiological relevance of the ligand-independent activation of PRs by DA. i.c.v. Administration of the DA receptor stimulant, apomorphine, or a D₁ agonist, SKF38393, facilitated receptive behaviour in oestradiol benzoate (EB)-primed female rats (21), confirming the earlier studies in which DA agonists were infused directly into the hypothalamus and preoptic areas of the rat brain (46). Interestingly, D₁ agonist facilitation of receptive behaviour in EB-primed female rats was inhibited by icv administration of PR antagonists, antisense oligonucleotides to PR mRNA or D₁ receptor antagonist (21), suggesting the existence of cross-talk between PRs and DA. The inability of EB-primed PR knockout mice to exhibit D₁ agonist-facilitated sexual behaviour, whereas their wild-type littermates responded to the agonist and exhibited lordosis responses (44), provides further evidence for the critical role of PRs as transcriptional integrators for the DA-dependent membrane-initiated pathways in female sexual behaviour.

Furthermore, our studies using antisense oligonucleotides to DA receptor subtypes indicate that DA-facilitated, PR-mediated behavioural effects occur via a D₁B (D₅) subtype and not the D₁A subtype (47). Using in situ hybridisation histochemistry, coexpression of D₁A/D₁B and PRs has been demonstrated in the medial preoptic area, lateral ventromedial nucleus of the hypothalamus and the arcuate nucleus of female rats (48).

Dopamine signalling and progestin receptor pathway convergence

In the mammalian brain, a tissue having an abundance of kinases and phosphatases, protein kinases and protein phosphatases plays an important role in the phosphorylation of signalling molecules involved in signal transduction mechanisms (49, 50). Neuronal phosphoproteins, like neurotransmitters and cyclic nucleotides, are components of the signal transduction pathway (51–53) and can be phosphorylated/dephosphorylated in response to extracellular stimuli. Such dynamic covalent modification is evident in modulation of the activity of PP1 and PP2. In the dopaminergic neurones, DA stimulates the phosphorylation of the phosphoprotein, dopamine and cAMP-regulated phosphoprotein-32 (DARPP-32) through a cascade of events involving the activation of D₁-subtype dopamine receptors, an increase in cAMP level, and activation of PKA (54). Phosphorylation by PKA on the Thr34 residue of DARPP-32 results in its conversion into a potent inhibitor of PP1 (54, 55). PP1 has broad substrate specificity and controls the state of phosphorylation and activity of numerous physiologically important substrates, including transcription factors, ion pumps, voltage-gated ion channels and neurotransmitter receptors. Thus, compounds that increase or decrease phospho-Thr34 of DARPP-32 inhibit or activate PP1 respectively, thereby increasing or decreasing the state of phosphorylation and activity of a large array of downstream physiological effectors (51). PR is one of the potential substrate proteins activated by DARPP-32.

Antisense oligonucleotides to DARPP-32 administered i.c.v. into the third cerebral ventricle inhibited D₁ agonist- and progesterone-facilitated sexual receptivity in E₂-primed female rats (56). D₁agonist- and progesterone-facilitated sexual receptivity also was inhibited in E₂-primed female mice carrying a null mutation for DARPP-32 gene. Similar to the effects of DA in the neostriatum, D₁ agonist, as well as progesterone, significantly increased hypothalamic cAMP levels and PKA activities, and enhanced the phosphorylation of DARPP-32 on Thr34 (56). D₁ agonist-induced increases were inhibited by the D₁ subclass DA antagonist, SCH 23390, indicating that the increases were due to the effects of DA initiated at its membrane receptor (56). Progesterone-induced increases, however, were not inhibited by SCH 23390, suggesting that the observed increases were due to the direct effects of progesterone and not secondary to modulation of DA receptors by progesterone (56). Rp-cAMPS, a compound that blocks cAMP signal transduction cascade by inhibiting PKA, inhibited D₁ agonist- and progesterone-facilitated sexual receptivity in E₂-primed female rats (56). These observations indicate that DARPP-32 activation is an obligatory step in PR regulation of sexual receptivity. It is likely that the mechanisms include both modulation of DARPP-32, making it an efficient inhibitor of PP1 (51) with an IC₅₀ and K_i of approximately 10⁻⁹ m (55), as described below, and/or the effects of DARPP-32 on the phosphorylation (activation) of PRs and/or PR-associated coactivators (57). The sequence of downstream events leading to the inhibition of PP1 and activation of PR (from DARPP-32 phosphorylation step) is currently under investigation.

Regulation of PP1 cascade by multi-site phosphorylation of DARPP-32

DARPP-32, a 205 amino acid protein is highly conserved in mammals. Recent studies on DARPP-32 regulation in dopaminoceptive neurones have identified multiple phosphorylation sites on DARPP-32 at Thr75, Ser102 and Ser137, in addition to Thr34 (51). This multi-site phosphorylation involves enzyme-directed and substrate-directed complex feedback loops and amplifies the effects of DARPP-32 by converting it into a better substrate for phosphorylation at Thr34, by PKA, contributing to its inhibitory effects on the downstream PP1 cascade (51) (Fig. 1).

Figure 1.  Multisite phosphorylation of the phosphoprotein, dopamine and cAMP-regulated phosphoprotein (DARPP)-32. DARPP-32 is phosphorylated at Thr34 by protein kinase (PKA), at Thr75 by cyclin-dependent kinase 5 (Cdk5), at Ser102 by casein kinase (CK)2 and at Ser137 by CK1. Phospho-Thr34 is preferentially dephosphorylated by protein phosphatase (PP)2B (Calcineurin); phospho-Thr75 is dephosphorylated by PP2A; phospho-Ser137 is dephosphorylated by PP2C; the phosphatase that dephosphorylates Ser102 is unclear. Phosphorylation of Thr34, by PKA, converts DARPP-32 into a potent inhibitor of PP1. Phosphorylation at Thr75 converts DARPP-32 into an inhibitor of PKA, reducing its ability to phosphorylate DARPP-32 and other substrates (filled arrows indicate positive effects; empty arrows indicate a negative effect). Phosphorylation of Ser102 and Ser137 acts to increase the phosphorylation of Thr34, by activating PKA and inhibiting dephosphorylation of T34 by PP2B, respectively. This phosphorylation potentiates dopamine (DA) and/or progesterone (P) signalling via the cAMP/PKA/DARPP-32/PP1 pathway. CK, casein kinase; PR, progestin receptor.

DARPP-32 is a dual function protein acting as an inhibitor of PP1 or of PKA, depending on whether it is phosphorylated on Thr34 or Thr75 (58). Phosphorylation at Thr34, by PKA, converts DARPP-32 into a potent inhibitor of protein PP1 (59). Phosphorylation at Thr75, by cyclin-dependent kinase 5, switches DARPP-32 into an inhibitor of PKA, reducing its ability to phosphorylate any substrate including DARPP-32 on Thr34 (58, 60). Activation of D₁ receptors. Results in increased PKA activity and phosphorylates Thr34, leading to the inhibition of PP1, and a decrease in the dephosphorylation of various physiological substrates. The activated PKA in turn activates another phosphatase, PP2A, in the dopaminergic neurones, which causes the dephosphorylation of Thr75, thereby removing the PKA inhibition. Thus, the PKA/PP2A/Thr75 DARPP-32 triad amplifies the effects of dopamine on the Thr34-DARPP-32/PP1 cascade (Fig. 1). Although the involvement of this feedback loop on DA-mediated DARPP-32 regulation has been extensively studied in the neostriatum and nucleus accumbens, its role in progesterone-regulation of the PR-sensitive areas of the hypothalamus is unknown. More importantly, the regulation of this loop by progesterone and its effects on progesterone-facilitated receptive behaviours are currently under investigation in our laboratory.

DARPP-32 is efficiently dephosphorylated by PP2B/calcineurin (61, 62). Ser102 and Ser137 phosphorylation on DARPP-32 modulate this dephosphorylation. In intact neurones, DARPP-32 is highly phosphorylated on Ser102 and Ser137. Phosphorylation at Ser102, by casein kinase (CK)2, enhances the ability of PKA to phosphorylate DARPP-32 at Thr34 (63), whereas phosphorylation at Ser137, by CK1, inhibits dephosphorylation at Thr34 by protein phosphatase 2B (64, 65). Thus, the physiological effect of Ser102 and Ser137 phosphorylation events potentiate signalling through the PKA/DARPP-32/PP 1 pathway and reduce signalling through the dephosphorylation of DARPP-32 on Thr34. Interestingly, the enzymes CKI and CKII, which phosphorylate DARPP-32 in vivo and in vitro (63–65), are also known to regulate the phosphorylation of PR (66). Thus, phosphorylation of serine residues on DARPP-32 could be critical for progesterone-facilitated signals in female reproductive behaviour.

We have recently begun to examine the effects of this intricate regulation of DARPP-32 in progesterone- and DA-facilitation of behaviours using mutant mice with point mutations of Thr34, Thr75, Ser102 and Ser137 (67) and such studies are ongoing. Such a complex regulation of DARPP-32 at multiple sites could provide a mechanism by which signal amplification is achieved to selectively alter the effects of progesterone and DA on behaviour and physiology.

Conclusions

It is becoming abundantly clear that the integration of the ligand-dependent and ligand-independent mechanisms of PR activation is essential for neuroendocrine regulation of complex behaviours. The cellular and molecular mechanisms involved in the convergence indicate substantial ‘cross-talk’ between signal transduction cascades initiated at the membrane leading to activation of intracellular PRs and the behavioural response. Neuronal kinases and phosphatases also play a predominant role in the regulation of the equilibrium between transcriptionally active and inactive states of PRs and their coregulators. This equilibrium could be fine-tuned by neuronal phosphoproteins (e.g. DARPP-32) functioning as signal amplifiers. Mutual interdependence and convergence of signalling pathways appear to be a ‘reinforcing’ mechanism by which neuronal responses to environmental, behavioural events alter the effects of steroid hormones on behaviour and physiology.

Acknowledgements

This work was supported by the United States Public Health Service grants MH57442 and MH63954 (S.K.M.).