Oestrogen Receptors and Signalling Pathways: Implications for Neuroprotective Effects of Sex Steroids in Parkinson’s Disease

Authors


Dr Thérèse Di Paolo, Molecular Endocrinology and Genomic Research Center, CHUQ-CHUL, 2705 Laurier Boulevard, Quebec City, QC, Canada G1V 4G2 (e-mail: therese.dipaolo@crchul.ulaval.ca).

Abstract

Parkinson’s disease (PD) is an age-related neurodegenerative disorder with a higher incidence in the male population. In the 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) mouse model of PD, 17β-oestradiol but not androgens were shown to protect dopamine (DA) neurones. We report that oestrogen receptors (ER)α and β distinctly contribute to neuroprotection against MPTP toxicity, as revealed by examining the membrane DA transporter (DAT), the vesicular monoamine transporter 2 (VMAT2) and tyrosine hyroxylase in ER wild-type (WT) and knockout (ERKO) C57Bl/6 male mice. Intact ERKOβ mice had lower levels of striatal DAT and VMAT2, whereas ERKOα mice were the most sensitive to MPTP toxicity compared to WT and ERKOβ mice and had the highest levels of plasma androgens. In both ERKO mice groups, treatment with 17β-oestradiol did not provide neuroprotection against MPTP, despite elevated plasma 17β-oestradiol levels. Next, the recently described membrane G protein-coupled oestrogen receptor (GPER1) was examined in female Macaca fascicularis monkeys and mice. GPER1 levels were increased in the caudate nucleus and the putamen of MPTP-monkeys and in the male mouse striatum lesioned with methamphetamine or MPTP. Moreover, neuroprotective mechanisms in response to oestrogens transmit via Akt/glycogen synthase kinase-3 (GSK3) signalling. The intact and lesioned striata of 17β-oestradiol treated monkeys, similar to that of mice, had increased levels of pAkt (Ser 473)/βIII-tubulin, pGSK3 (Ser 9)/βIII-tubulin and Akt/βIII-tubulin. Hence, ERα, ERβ and GPER1 activation by oestrogens is imperative in the modulation of ER signalling and serves as a basis for evaluating nigrostriatal neuroprotection.

Introduction

Numerous studies (1–7), a meta-analysis (8) and reviews (9, 10) indicate that Parkinson’s disease (PD) is more prevalent and has a larger incidence in the male population. A meta-analysis of seven studies that used a stringent inclusion criteria reported that overall PD is 1.5-fold greater in men than in women (8). The possible reasons proposed for this increased risk of PD in men than in women include toxicant exposure, head trauma, neuroprotection by oestrogens, mitochondrial dysfunction and/or X linkage genetic risk factors (8).

Gender differences were reported during the progression of PD and in the responses to l-dopa treatment (11–13). In the case of PD, oestrogens have been known to protect the nigrostriatal dopaminergic pathway (14, 15). However, the beneficial effects of oestrogens remain far from simple; complexity is underlined by their various neuroprotective actions and numerous mechanisms involved. Hence, research continues concerning these mechanisms and the development of new or improved neuroprotective compounds. We propose that oestrogenic neuroprotection arises through oestrogen receptors (ERs), is influenced by their subtype and we review their implication in neuroprotection as derived from results from our laboratory using lesioned primates and mice models of PD, as well as the current literature.

PD

Characteristics, cause and symptoms of PD

PD is a neurodegenerative disorder that has become more prevalent over the years as a result of an ageing population (16). This disease has no documented aetiology; however, it is characterised as chronic and progressive, with the loss of dopamine (DA)-containing neurones in the brain substantia nigra (SN) being a main feature in its neurodegenerative process (16). The disorder is also referred to as ‘idiopathic’ PD because no known cause is documented, yet some cases may be a result of toxicity, drugs, genetic mutation, head trauma or other medical disorders (16). PD is also characterised as ‘sporadic’ with no known genetic background in 90–95% of cases, and environmental factors or genetic susceptibilities are considered to have a hand in triggering the disease (16).

What makes PD so infamous is the fact that symptoms appear after a myriad of neurones are lost, with a death toll of dopaminergic neurones exceeding a critical threshold of between 70% and 80% in the SN (16). It is hard to diagnose before the damage becomes irreversible as a result of compensatory mechanisms (16). Disruption in DA transporters that are responsible for controlling DA concentrations, and include the DA transporter (DAT) and the vesicular monoamine transporter 2 (VMAT2), are considered to play a role in this pathology (17).

The 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) model of PD

MPTP provides a model of PD as a result of its specific neurotoxicity in certain species such as mice and monkeys (18, 19). It was discovered as a by-product in the chemical synthesis of a meperidine analog with potent heroin-like effects (19). The irreversible effects of MPTP in man induce tremor, slowness of movement, rigidity, freezing and postural instability, thus mimicking almost all features of PD (19). MPTP enters the brain through the presynaptic DA uptake system where it is converted into the 1-methyl-4-phenylpyridinium (MPP+) ion by monoamine oxidase B (19). MPP+ is a very potent neurotoxin that inhibits the mitochondrial oxidation of NAD+-linked substrates. It accumulates inside mitochondrial matrixes and inhibits respiration (19), resulting in the death of nigrostriatal nerves through mechanisms that include the loss of calcium homeostasis and the formation of radicals that promote cytotoxic events leading to limited DA release, locomotor defects and finally apoptosis (19).

Oestrogens and PD

Several studies indicate that PD has a later age at onset in women and that men are more susceptible (15). Clinical studies demonstrate that there is an increased risk of developing PD under conditions of an early decrease in natural endogenous oestrogens (20, 21).

A 2-week double-blind cross-over study, on postmenopausal women suffering from mild to moderate PD showed that transdermal treatment with a high dose of 17β-oestradiol reduces the antiparkinsonian threshold dose of levodopa (22). Oestrogens were observed to have modulatory effects on PD symptoms and levodopa-induced dyskinesias (23–25) and motor disabilities were remedied by oestrogens in postmenopausal women with PD (26). On the basis of evidence from a case–control design study of 133 female PD cases and 128 female controls, it was concluded that women who took a postmenopausal oestrogen treatment had a lower risk of developing PD (27). Men with PD show symptoms requiring medical attention during earlier stages of the disorder (21), suggesting that the disease progresses more rapidly in men, thus supporting the idea that oestrogen can provide neuroprotective effects (21). Furthermore, gender differences in symptoms were also seen in outcome studies after stereotactic surgery for PD (10).

ERs

Classic ERs

Currently, there are two main documented ER subtypes (28). The transcriptional information for these two receptors lies on two different genes (28). Twenty splice variants have been documented for ERα and ten for ERβ (28). There is little or no conservation in their N-terminal region, although they have a homology of 55% in their ligand-binding domains and are 95% homologous in their DNA-binding domains (28). Oestrogens may differ in their affinity for either ER subtype. For example, 17β-oestradiol binds equally with both ERα and ERβ, whereas oestrone and the selective oestrogen receptor modulator, raloxifene, preferentially bind ERα and oestriol has more affinity for ERβ (29). Each ER has its own nonsteroidal agonist (30, 31); 1,3,5-tris(4-hydroxyphenyl)-4-propyl-1H-pyrazole (PPT) is selective for ERα and 2,3-bis(4-hyroxyphenyl) propionitrile (DPN) is selective for ERβ.

ERα is present in high levels within the cardiovascular system, bone, mammary glands and the uterus, whereas ERβ is found mainly in the urinary tract, prostate and ovaries (32, 33). At the cellular level, the nucleus is mainly home to ERα, whereas ERβ is reported to be localised in the cytoplasm (34). In the central nervous system (CNS), these two receptors have similar patterns of expression in the preoptic area, the cortical amygdaloid nuclei and the bed nucleus of the stria terminalis. In rodents, ERα expression is found exclusively in the ventromedial hypothalamic nucleus and the subfornical organ (35) and it is predominant in mouse hippocampus (36), whereas ERβ is predominant in rat hippocampus and the cerebral cortex (35). Immunohistochemical and in situ hybridisation data show that ERα and ERβ are present in rat SN (37). In addition, Kritzer et al. (38) showed the colocalisation of ERs and tyrosine hyroxylase (TH) in the ventral tegmental area and SN of the rat brain and subsequently specified that ERβ is present on collateral SN pars compacta projections to the ventral striatum (39). Küppers et al. (40) claim that both ERα and ERβ mRNAs are expressed in GABAergic neurones of the striatum.

Nuclear ER activity

ERs regulate gene transcription by exerting positive or negative effects on the expression of target genes. ERα and ERβ are ligand-activated receptors that work with the help of two DNA consensus elements (41). These receptors regulate gene transcription and so they are part of a vast family of proteins called ligand-activated transcription factors. Their actions have been documented to include either genomic or nongenomic mechanisms and the fact that both ERs are found in dopaminergic neurones of the midbrain, although in low abundance, implies that the adult nigrostriatal system is targeted by oestrogens (40, 42). The genomic mechanism could take two different routes: either direct or indirect. When not bound to a ligand, ERs are found as monomers that associate with heat shock protein (Hsp-90) and immunophilins forming a multiprotein complex (43). Once oestrogen binds a complementary receptor, the phosphorylation of its many different serine/threonine residues is induced, which causes them to lose the Hsp-90 and change their conformation to promote their homo- or heterodimerisation and translocation into the nucleus. A hydrophobic clef is also revealed to bind transcriptional coactivators that help by initiating a chromatin structural change in target promoters (44, 45). Once inside, ERs interact with oestrogen response elements (EREs) on the regulatory sequences of target genes to either suppress or activate their transcription at the same time as being limited to both promoter and cell specificities (46). However, they can also function without EREs with the help of ER-tethering and coactivation of transcription factors bound to the target DNA, such as the transcriptional factor cAMP-response element binding protein. They can also interact with fos/jun transcription factors were they can regulate transcription through activator protein-1 (44, 47–49).

The indirect route is initiated by disinhibiting mitogen-activated protein kinases (MAPK) (50, 51) and protein kinase B (PKB/Akt) signalling cascades (52), thus leading to the activation of cAMP-response element binding protein (53, 54). Ultimately, the modulation of proteins such as Bcl-2 and Bad that regulate apoptosis is induced.

Membrane ERs

Genomic mechanisms require hours to run their course because they involve the transcription and translation of oestrogen-regulated genes (55). However, not all oestrogenic effects are attributable to nuclear ERs; some effects occur within a matter of seconds or minutes, making it obvious that a route other than one implicating classic intranuclear receptor transcription modulation has transpired. Membrane-associated ERs could provide the basis for these observations. This theory gained support over the past 30 years ever since Pietras and Szego (56) first observed receptors as having rapid responses to 17β-oestradiol and continues today with evidence of their existence as a result of the fast nongenomic effects of 17β-oestradiol. Although ERα and ERβ are both able to act as plasma membrane receptors (57), there is also evidence for new plasma ERs categorised as G protein-coupled receptors (GPCRs), which have no relation to the known ERs (58). How ERs are brought to the plasma membrane is not well established, although research on non-neuronal cells supports the theory that ER palmitoylation could be a mechanism (59, 60). Here, a post-translational addition of a 16-carbon fatty acid (palmitate) to ER ligand-binding domain residues enables targeting of ERs to the plasma membrane (59, 61).

The cloning of a G protein-coupled ER named GPER1 (also known as GPR30) was reported over 15 years ago in several studies (62–67). GPER1 is distinct from the classic ERα and ERβ and has been shown by electron microscopy to be located on the plasma membrane (68). Its expression in the rat CNS, as determined through immunohistological studies, is high in brain regions, including the cortex, striatum, hippocampus, SN, islands of Calleja, hypothalamic-pituitary axis and the brainstem autonomic nuclei (69), whereas its cellular localisation also includes the endoplasmic reticulum and the Golgi apparatus (70, 71). GPER1 has a high affinity for 17β-oestradiol but 17α-oestradiol failed to significantly displace [3H]oestradiol binding to GPER1 (70, 71). The competitive binding assay shows that GPER1 does not bind cortisol, progesterone or testosterone, whereas oestrone, oestriol had very low affinities for the receptor (72); tamoxifen, genistein and ICI 182,780 are agonists on GPER1 (73). Hammond et al. (74) observed that GPER1 is expressed (0.4–42%) in GABAergic neurones and that it is colocalised in 63–99% of cholinergic neurones in the basal forebrain. However, Mufson et al. (75) and Shughrue et al. (76) observed that approximately 30% of the cholinergic neurones also contain ERα. In mice, another membrane ER, named ER-X, was found to be expressed in the neocortex, in lung plasma membrane microdomains associated with caveolin, and also in the uterus (58, 77, 78).

Oestrogens, ERs and neuroprotection

It is well established that ageing results in a decline in the production of the female hormone oestrogen, with the most drastic decrease at menopause (79). In addition, it has been observed that, with ageing, negative effects on ERα-mediated events are assumed to attenuate ER functioning by increased methylation of the ER gene (80, 81). In age-related animal models of neurodegenerative diseases, numerous studies show that oestrogens play an important protective role and there is accumulating evidence implicating ERs.

It is expected that nuclear ER concentrations in man diminish with ageing, as reported in rats (82). However, in humans, post-mortem tests show complex age-related alterations of the canonical ERα and various ERα splice variants in the brain and the expression pattern of certain forms is brain area-specific (83). In the vasopresssinergic supraoptic nucleus and the hippocampus, ERα was reported to increase with advancing age in women with higher expression in postmenopausal than in pre- and perimenopausal women (84). ERα splice variant del.7 (deletion of exon 7) and del.2 (deletion of exon 2) declined with advancing age (61–84 years old) in the mamillary body but not in the hippocampus (83), whereas no change was observed in another study of 12 exon-skipping variants, with the most common form found being del.5, del.7 and del.2 in people aged 29–59 years (85). The del.7 is a dominant negative variant that can inhibit transcriptional activity of both ERα and ERβ by forming heterodimers (86). del.4 was observed in the caudate nucleus, putamen and SN of a 71-year-old female Alzheimer's disease patient but the effect of ageing was not reported (83). del.4 is suggested to be a silent variant without activity of its own, although a dominant negative function was ascribed to this slice form through protein-protein interactions with ERα (87). Hence, higher levels of ERs with hindered activity are seen in some brain regions as a function of ageing (88).

The effects of MPTP on dopaminergic markers

Vesicular monoamine transporter 2 packages serotonin, histamine, epinephrine, norepinephrine and DA into vesicles and is mainly confined to the CNS (89). The DAT is a specific protein of DA neurones (90). DAT and VMAT2-specific binding under pathological conditions is used as a marker to evaluate DA cell body and terminal integrity. Neurotoxicity of SN DAT-specific binding and mRNA are less severe than that seen in the striatum of MPTP and methamphetamine lesioned mice (14, 91), suggesting that the presynaptic DA terminals of the striatum are more vulnerable. Accordingly, high doses of neurotoxins are needed to affect SN TH mRNA levels (92). Indeed, cell body destruction requires elevated doses of toxins, whereas cell terminals are lost with lower doses of the same neurotoxins (92). Hence, different stages exist where neurones are injured but not dying or dead when neuroprotection is feasible by steroids and, in humans, oestrogens are only beneficial before starting levodopa therapy in the early stages of PD (93).

Nigrostriatal DA activity regulation depends on DA availability, which in turn is controlled by DAT and VMAT2 present in neurones. A role for DAT in astrocytes is also reported because Karakaya et al. (94) observed that DAT is expressed in neonatal astrocytes and that 17β-oestradiol dose-dependently down-regulated DAT mRNA by 80% and 60% in the neonatal midbrain and striatal astroglia cultures, respectively. It was also noted that 17β-oestradiol inhibits the clearance of extracellular DA by 45% and 35% in the neonatal midbrain and striatal astroglia cultures; this effect was abolished with the use of an ER antagonist ICI 182, 780 (94). It was concluded that the effects of 17β-oestradiol on DAT could be neuroprotective under pathological conditions because the end result is delayed DA uptake by astroglia (94). Thus, the 17β-oestradiol-induced decrease of astroglial DA uptake may diminish DA metabolisation, resulting in an increased availability of synaptic DA and, subsequently, more DA for recycling by neurones. VMAT2 was not detected in astrocytes (94).

Oestradiol neuromodulation and neuroprotection

Oestrogens modulate the nigrostriatal and mesolimbic DA systems’ activity at various components of neurotransmission (95–97). We have shown, by biochemical and pharmacological studies, that 17β-oestradiol can modulate DA receptors (98) and DAT (99–101). Striatal and nucleus accumbens DA D2 receptor and DAT density are increased with chronic 17β-oestradiol treatment without affecting their mRNA levels, implying that 17β-oestradiol activity was nongenomic (102, 103).

Oestrogens produce their modulatory effects through pro- or anti-dopaminergic activity, such as on enzymes that synthesise or degrade DA, DAT, VMAT2, DA receptors and DA release (15). In the striatum of ovariectomised rats, DAT density fluctuations are observed during the oestrous cycle (97) and 17β-oestradiol treatment increased DAT density through acute and chronic treatment (101, 104), whereas reductions in SN DAT mRNA levels in ovariectomised rats were restored with oestrogens (99).

Postmenopausal women given an oestrogen replacement therapy had increased DAT density in the left anterior putamen (105). There is less data available on the gonadal hormone modulation of VMAT2; in the rat brain, 17β-oestradiol treatment had no effect on its striatal density (106) and, in another study in the SN pars compacta, chronic 17β-oestradiol treatment did not affect VMAT2 mRNA (107).

Oestrogens provide relief from PD symptoms if treatment is given at early stages of the disease (93). Hence, to model this stage of PD, we used conditions of moderate nigrostriatal DA loss in MPTP mice when motor behaviour is not yet impaired, nor is there significant DA cell death. Thus, under conditions of early nigrostriatal DA neuronal degeneration, MPTP mice show a significant reduction of striatal DA concentrations and DA transporter loss. The neuroprotective effects of 17β-oestradiol, when administered before the MPTP regimen, are observed upon the prevention of DA and dopaminergic metabolite depletions (108–111), as well as DAT and VMAT2-specific binding loss (112). In addition, 17β-oestradiol treatment promotes an increase in TH immunoreactive neurones of the SN pars compacta of male mice (113). The neuroprotective effects of 17β-oestradiol upon MPTP are achieved with pretreatment at low doses mimicking physiological levels in male and female mice, although treatment with high doses does not prevent MPTP neurotoxicity (114). Oestrogens are considered to convey their neuroprotective effects through genomic mechanisms that signal through ERs or by using nongenomic mechanisms through membrane bound receptors (115–117). Again, we propose that these neuroprotective effects are mediated via an interaction with ERs. This is supported by the observation that 17α-oestradiol, which has a low ER affinity, does not induce neuroprotective effects (91). Moreover, the weak ER agonists, oestrone and oestriol (91), have weak or no neuroprotective potentials against dopaminergic loss caused by MPTP (118).

ERα and ERβ could have distinct roles in neuroprotection against MPTP toxicity (119), with ERα being the dominant receptor involved in neuroprotection (120, 121). Indeed, it is speculated that ERβ plays a less dominant role in neuroprotection because its activity is optimal once cellular death is inhibited and regeneration and neurogenesis commence (122). Dubal et al. (120) show that, in a cerebral ischaemia model, 17β-oestradiol treatment does not protect the cortex or striatum in ER knockout (ERKO)α mice compared to wild-type (WT) and ERKOβ mice. Our group has demonstrated that PPT but not DPN provides neuroprotection against MPTP (123, 124). Similarly, PPT but not DPN protects against β-amyloid peptide in cerebrocortical neuronal cultures via a protein kinase C-dependent signalling pathway (125). Other studies claim that ERβ plays a role in neuroprotection. For example, Carswell et al. (126) provide data suggesting that DPN but not PPT pretreatment reduces ischaemic damage in the striatum and CA1 region of the hippocampus and Westberg et al. (127) report that ERβ gene polymorphisms could influence the age of onset of PD. We reported that treatment with DPN or 17β-oestradiol but not PPT modulates D2 DA receptors in ovariectomised rats (102), whereas data obtained from a study with ERKOα and ERKOβ male mice suggest that ERβ affects DA metabolism because ERKOβ mice had a lower DA turnover rate (119). 17β-Oestradiol was able to prevent the loss of 3β-(4-125I-iodophenyl)trophane-2β-carboxylic acid binding to DAT and [3H]-dihydrotetrabenazine binding to VMAT2 in the striatum and SN of MPTP and methamphetamine lesioned mice (114). Moreover, the SN decrease of these transporters’ mRNA produced by MPTP was prevented by oestrogen (91). We have also examined the neuroprotective contributions of ERα and ERβ against MPTP toxicity by examining DAT, VMAT2 and TH in ERKO C57Bl/6 male mice.

Our results show striatal DAT and VMAT2 levels of intact ERKOβ mice to be lower than WT and ERKOα mice, whereas ERKOα had elevated plasma androgen concentrations (two-way anova shows an effect of genotype, P < 0.0001; mean ± SEM, testosterone: WT = 3.5 ± 0.9 ng/ml, ERKOα = 12.5 ± 1.0 ng/ml, P < 0.00001, versus WT and ERKOβ = 6.1 ± 1.2 ng/ml; dihydrotestosterone: WT = 114 ± 32 pg/ml, ERKOα = 997 ± 117 pg/ml, P < 0.0001, versus WT and ERKOβ = 149 ± 59 pg/ml) (128). This is in agreement with a previous report showing a significant increase of testosterone in ERKOα male mice compared to WT males (129). Despite being infertile, ERKOα mice have a close to normal hormonal profile and activity of the hypothalamic/pituitary axis (130), whereas ERKOβ mice have been found to be in a state of systemic hypoxia (131). Functional alterations of the SN DA system, as well as reduced TH and brain-derived neurotrophic factor levels, are observed in both male and female ERKOα mice (132). ERKO could affect the maturation of other components of brain DA transmission, such as the DAT and VMAT2, although this possible effect has not been verified.

In WT and ERKO mice, MPTP caused a dose-dependent loss of both striatal transporters (Fig. 1) that correlated with their previously reported reductions in striatal DA concentrations (119) (DAT: R = 0.755, P < 0.0001; VMAT2: R = 0.787, P < 0.0001). Compared to WT and ERKOβ, DAT, VMAT2 and TH showed a greater sensitivity to MPTP in ERKOα mice (Fig. 1, and data not shown). WT mice were compared with ERKO mice pretreated with 17β-oestradiol alone and/or with an effective dose of MPTP. The striatum and SN of ERKOα mice were more vulnerable to MPTP toxicity and 17β-oestradiol protected against this toxicity only in WT mice (Fig. 2, and data not shown) despite similar plasma 17β-oestradiol concentrations among the three genotypes (two-way anova shows an effect of 17β-oestradiol treatment, P < 0.0001; mean ± SEM in pg/ml, WT: vehicle: 2.2 ± 1.5 and 17β-oestradiol treated: 10.0 ± 1.4; ERKOα: vehicle: 4.3 ± 1.4 and 17β-oestradiol treated: 11.2 ± 2.0; ERKOβ: vehicle: 3.3 ± 0.8 and 17β-oestradiol treated: 10.2 ± 1.2). Hence, even though the lack of the ERα caused a more significant susceptibility to MPTP toxicity, both ERα and ERβ were shown to be implicated in neuroprotection resulting from 17β-oestradiol.

Figure 1.

 Dose-response effect of 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) on striatal dopamine transporter (DAT) and vesicular monoamine transporter 2 (VMAT2)-specific binding measured with 3β-(4-125I-iodophenyl)trophane-2β-carboxylic acid ([125I]-RTI-121) and [3H]-dihydrotetrabenazine ([3H]-TBZ-OH) binding respectively in wild-type (WT) compared to oestrogen receptor knockout (ERKO)α and ERKOβ mice using our assay conditions (104, 165). Experimental details of treatments of these mice and their striatal biogenic amine concentrations were described previously (119). DAT: F11,57 = 20.3, P < 0.0001 and VMAT2: F11,55 = 30.6, P < 0.0001; *P < 0.05, **P < 0.01, ***P < 0.005 and ****P < 0.0001 versus respective intact, vehicle (saline/gelatine solution, 0 MPTP); ††P < 0.01, †††P < 0.005 and ††††P < 0.0001 versus WT MPTP 7 mg/kg; ‡‡‡‡P < 0.0001 versus WT MPTP 9 mg/kg; £P < 0.05, £££P < 0.005 and ££££P < 0.0001 versus WT MPTP 11 mg/kg; &P < 0.05 and &&&P < 0.005 versus WT vehicle (0 MPTP); ΦP <0.05, and ΦΦΦP < 0.005 versus respective experimental ERKOα genotype group.

Figure 2.

 Effect of treatment with 17β-oestradiol (E2) (2 μg/day) for 10 days on anterior striatal dopamine transporter (DAT) and vesicular monoamine transporter 2 (VMAT2)-specific binding, measured with 3β-(4-125I-iodophenyl)trophane-2β-carboxylic acid ([125I]-RTI-121) and [3H]-dihydrotetrabenazine ([3H]-TBZ-OH), respectively, in intact and 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) (9 mg/kg) lesioned wild-type (WT), oestrogen receptor knockout (ERKO)α and ERKOβ mice. For experimental details, see Fig. 1. DAT: F11,69 = 32.7, P < 0.0001 and VMAT2: F11,68 = 31.2, P < 0.0001; *P < 0.05, **P < 0.01, ***P < 0.005, and ****P < 0.0001 versus respective intact, vehicle (0 MPTP); P < 0.05 and ‡‡‡‡P < 0.0001 versus WT MPTP; P < 0.05 and •••P < 0.005 versus WT + E2; ++++P < 0.0001 versus WT + MPTP + E2; ΦP < 0.05 and ΦΦΦΦP < 0.0001 versus respective ERKOα genotype group.

The absence of ERα and ERβ throughout development in ERKO mice could affect brain organisation and may result in a different adult brain. However, our results on WT mice showing that PPT but not DPN protects striatal DA against MPTP toxicity (123, 124) are in agreement with our results demonstrating a greater susceptibility of the nigrostriatal DA system of ERKOα mice to MPTP (128). Hence, the ERKO mouse model is a valid tool for the study of the role of ERs in nigrostriatal DA neuroprotection and supports the results obtained using ER-specific agonists.

Neuroprotective implication of membrane ERs

Each ER has many splice variants (133) and research has shown that ERα and ERβ are able to act as plasma membrane receptors (133–135). There is also evidence of new plasma ERs categorised as neither ERα, nor ERβ (136, 137). These so-called, GPCRs have no relation to the known ERs (138). However, Marin et al. (134) conclude that membrane ERα and ERβ are homologues of the nuclear ERs. If correct, then membrane ERα and ERβ should be absent in ERKOα and ERKOβ. There is evidence that intracellular ERα and ERβ are transported to the membrane and their interaction with metabotropic glutamate receptors provides many possibilities for membrane associated 17β-oestradiol cell signalling mediation (139). On the cellular membrane, ERα and ERβ activity resembles that of GPCRs and oestradiol modulates membrane-associated ERα and ERβ by inducing their internalisation (139). Although some controversies underlying its localisation and activity still exist (61), it is clear that GPER1 mediates rapid as well as transcriptional oestrogenic activity in the brain and periphery (73). GPER1 activity is manifested through two plasma membrane-associated enzymes; the first is Gs-protein, which induces adenylyl cyclase promoting elevated intracellular concentrations of cAMP, and the second is Gβγ, which results in calcium mobilisation and kinase activation (140).

However, the role of GPER1 in the brain DA systems and neurodegenerative disorders has yet to be determined. Potential tools for elaborating the physiological activities of this new ER in the brain include the GPER1-specific agonist, G1 and antagonist, G15, both of which exist without any activity on ERα or ERβ (141, 142). As noted above, a vast amount of literature provides evidence that oestrogens have positive effects on the DA system in the brain, and their classical mechanisms on nuclear ERs should be studied along with their potential GPER1 activity. Therefore, we aimed to assess the distribution of striatal GPER1 and its response to oestradiol and lesions. High levels of GPER1 were measured in the striatum of male mice and were increased by MPTP lesion (Fig. 3). Moreover, we compared methamphetamine-induced neurotoxicity on striatal GPER1 in male and female mice; methamphetamine, which produced significantly decreased striatal DA in males, increased striatal GPER1 levels in male, but not in female mice (143). Ovariectomised female Macaca fascicularis monkeys with a unilateral MPTP lesion of the nigrostriatal pathway that received a chronic 17β-oestradiol or vehicle treatment for 1 month were studied next. The lesioned striata of these monkeys were extensively denervated, as indicated by reductions in DA concentrations (144). GPER1 levels were abundant in the monkey striatum, both in the caudate nucleus and the putamen, at the two rostro-caudal levels measured. Similar levels of GPER1 were found in the anterior versus posterior caudate nucleus and putamen of monkeys (Fig. 4). GPER1 levels were higher in the putamen, but not the caudate nucleus, in the MPTP-lesioned side compared to that of the intact side of hemiparkinsonian monkeys. 17β-Oestradiol treatment did not significantly change GPER1 levels in the intact or lesioned caudate nucleus and putamen of monkeys (Fig. 4). The present results show that GPER1 is abundant in the striatum of monkeys and mice and is increased in response to toxins that target the nigrostriatal pathway.

Figure 3.

 Effect of 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) (11 mg/kg) in C57Bl/6 male mice on striatal dopamine (DA) and its metabolites, dihydroxyphenylacetic acid (DOPAC), homovanillic acid (HVA) and 3-methoxytyramine (3-MT) concentrations, as well as G protein-coupled oestrogen receptor (GPER1) levels. Catecholamines concentrations were measured under conditions previously reported by high-performance liquid chromatography with electrochemical detection (119) and GPER1 levels by western blotting (143). *P < 0.05, ***P < 0.005, and ****P < 0.0001 versus respective intact, vehicle (0 MPTP).

Figure 4.

 G protein-coupled oestrogen receptor (GPER1) levels in the caudate nucleus and putamen of hemiparkinsonian ovariectomised monkeys treated for 1 month with vehicle (n = 4) or with 17β-oestradiol (0.1 mg/kg once daily subcutaneous, n = 3). Data are expressed as a percentage of the intact side of vehicle-treated monkeys. Experimental details of treatments of these monkeys and their striatal biogenic amine concentrations were reported previously (144). GPER1 and βIII-tubulin were measured by western blotting under conditions described previously (143). *P < 0.05 and ****P < 0.0005 versus the intact side of respective treated monkeys.

Akt and GSK3 signalling pathways

Akt signalling in mental and neurodegenerative diseases

In addition to the classical functions of DA receptor-cAMP-dependent mechanisms, striatal DA D2 receptors can also exert physiological effects via PKB (Akt) (145). Akt can be activated after the binding of neurotransmitters or growth factors on many specific cell-surface receptors, which in turn initiate a cascade of second messengers related to the phosphatidylinositol 3-kinase (PI3-K) pathway (146). Deactivation of Akt after dephosphorylation results in activation of the glycogen synthase kinase-3 (GSK3)-mediated signal (147). Dysregulation of Akt/GSK3 signalling is involved in many DA-associated neurological and neuropsychiatric disorders. Reduced Akt activity or expression levels were shown in brains of schizophrenic patients (148, 149) and there are data indicating an involvement of GSK3β in depression and psychosis (150). A report showed an association between the Akt1 gene and PD that was a protective haplotype (151). In sections of post-mortem SN, an extensive reduction of pAkt(Thr308) and pAkt(Ser473) in PD patients was observed compared to controls (152).

The implication of PKB/Akt in 17β-oestradiol induced neuroprotection

The PI3-K/Akt and MAPK signalling pathways are associated with 17β-oestradiol activity in the brain (114). ERs relay MAPK signals through sequential activation of Ras, B-raf, MAPK/ERK kinase (MERK1/2) and MAPK (ERK1/2) to finally induce various transcriptional factors that promote neuronal survival (115, 135, 153, 154). The effector Akt can be activated via the PI3-K pathway through ERs (123). Akt activity promotes cell survival by modulating the expression of anti-apoptotic proteins such as Bcl-2 and apoptotic proteins like Bad and Bax (123) and signalling converges at GSK3β. GSK3β is a highly expressed kinase in the brain and, once activated, functions to induce cellular death pathways; therefore, its inactivation favours the promotion of cellular survival mechanisms (123). The activity of this kinase III is proapoptotic; it is inhibited if phosphorylated on certain serine residues, thus promoting cell survival (155) and activated if phosphorylated on tyrosine residues (115). Moreover, activation of these signalling pathways through either membrane-associated or genomic actions of 17β-oestradiol could be combined to act synergistically in injured neurones and amplify the neuroprotective process (156).

Data from several sources indicate that the MAPK pathway and ERβ contribute to cell survival signalling pathways in various models of neuronal injury (126, 157–159). However, Kahlert and colleagues have observed that ERα is involved in the activation of Akt signalling (160, 161), whereas ERβ is not (162). Our group, in collaboration with L. M. Garcia-Segura, reported that treatment with 17β-oestradiol increased phosphorylated PKB/Akt (at serine 473) levels in mice (123). We also observed that GSKβ3 phosphorylation on serine 9 was highly diminished in MPTP mice and that PPT treatment significantly blocked GSK3β activation, but not treatment with 17β-oestradiol or the ERβ agonist Δ5-diol (123). Only PPT treatment had positive effects by increasing the levels of inhibited GSK3β in MPTP treated mice and, in intact wild-type mice, it was the 17β-oestradiol and PPT treatments that increased the levels of activated PKB/Akt and deactivated GSK3β (123). These results support a role for the ERα in the PI3-K pathway being associated with the neuroprotective effects of oestrogenic compounds against MPTP. Moreover, GPER1 activated by 17β-oestradiol was also shown to initiate PI3-K signaling and Akt activity (70, 71).

No data are yet available on the oestrogenic modulation of this signalling pathway in monkeys. We thus measured the effect of 1 month of treatment with 17β-oestradiol on the Akt/GSK3 signalling pathway in the brain of ovariectomised monkeys with a unilateral MPTP lesion of the nigrostriatal pathway. 17β-Oestradiol treatment induced an increase of pAkt(Ser 473)/βIII-tubulin in the intact and lesioned posterior caudate nucleus and in pGSK3β(Ser 9)/βIII-tubulin in the intact and lesioned anterior putamen compared to vehicle-treated monkeys. In the intact and lesioned putamen, the Akt/βIII-tubulin was also increased in monkeys treated with 17β-oestradiol, whereas GSK3β/βIII-tubulin remained unchanged (Figs 5 and 6). These translational results in monkeys, similar to our previous findings in mice (123) suggest that activation of the Akt/GSK3 signalling pathway is involved in the 17β-oestradiol effect on the striatal DA system and support a beneficial role of oestrogenic treatment resulting from an increase in the activity of signalling pathways implicated in cell survival.

Figure 5.

 Relative levels of Akt and its phosphorylated form (pAkt, Ser473) in the caudate and putamen of hemiparkinsonian 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) monkeys treated for 1 month with vehicle or 17β-oestradiol. For experimental details, see Fig. 4. Akt and its phosphorylated form were measured by western blotting under conditions described previously (166). *P < 0.05 versus vehicle-treated monkeys; P < 0.05 versus the respective intact side.

Figure 6.

 Relative levels of phosphorylated glycogen synthase kinase-3β (GSK3β) (pGSK3β, Ser9) in the caudate and putamen of hemiparkinsonian 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) monkeys treated for 1 month with vehicle or 17β-oestradiol. For experimental details, see Fig. 4. GSK3β and its phosphorylated form were measured by western blotting as described previously (166). GSK3β/βIII-tubulin levels were unchanged by the lesion and 17β-oestradiol treatment (data not shown). *P < 0.05 versus vehicle-treated monkeys.

Concluding remarks

The combinations of the effects of nuclear ERα and ERβ and membrane GPER1 signalling pathways are reported to result in cell cycle progression and cell proliferation (73), which is beneficial in the injured or neurodegenerative brain. ERβ is reported to modulate ERα-mediated transcription in mice; therefore, in certain tissues and under certain conditions, these ERs are interdependent (163). Moreover, ERα and GPER1 were reported to cross-talk (73). There is great diversity in the possible synergistic or antagonistic interactions between ERs and GPER1 (73) and both membrane and fast transcription activity of GPER1 are reported to result in the activation of genes such as c-fos (164).

In summary, a complex cascade of genomic and nongenomic oestrogen-induced activity results in neuroprotection, which in turn relies on the neuroanatomical and spatio-temporal organisation of ERs and various signalling pathway molecules involved in their cross-talk in different neuronal populations. Oestrogens and oestrogenic drugs modulate and protect nigrostriatal DA activity and our results propose that ERs are implicated in these effects. A role for ERα and its agonists is observed in neuroprotection, whereas the novel GPER1 could provide a new target for modulating the nigrostriatal DA system. We also demonstrate that the Akt/GSK3 signalling pathway is modulated by oestrogens in intact and MPTP lesioned mice and monkeys. The implication of ERα, ERβ and GPER1 for modulation of nigrostriatal DA activity supports the development of a new generation of ER-specific drugs for the brain.

Acknowledgements

This work was supported by a grant from the Canadian Institutes of Health Research (CIHR) to T.D.P. S.A.S. and M.G.S. held a studentship from the Fonds de la recherche et de l’enseignement of the Faculty of Pharmacy of Laval University. M.B. holds a studentship from the Fonds de la Recherche en Santé du Québec (FRSQ).

Ancillary