|Home | About | Journals | Submit | Contact Us | Français|
Although many studies have focused on the neuroprotective effects of estrogen (E2) on stroke, there have been tantalizing reports that E2 may be neuroprotective in degenerative neuronal diseases such as Alzheimer’s and Parkinson’s disease. In animal models of Parkinson’s disease, E2 protects the nigrostriatal dopaminergic (DA) system against neurological toxins. However, little is known about the cellular and molecular mechanism(s) involved by which E2 elicits its neuroprotective effects on the nigrostriatal DA system. A preferred mechanism for neuroprotection is interaction of E2 with specific neuroprotective growth factors and receptors. One of such neuroprotective factor/receptor system is insulin-like growth factor-1 (IGF-1). E2 neuroprotective effects in the SN DA system have been shown to be dependent on IGF-1, a growth factor released by both astrocytes and neurons. To determine whether E2 also interacts with the insulin-like growth factor-1 receptor (IGF-1R) and to determine the cellular localization of estrogen receptor (ER) and IGF-1R we compared the distribution of ER and IGF-1R in the substantia nigra (SN). Stereological measurements revealed that a subpopulation (39.8%) of tyrosine hydroxylase-immunoreactive (TH-ir) SNpc DA neurons are immunoreactive to estrogen receptor-β (ERβ). No immunolabeling for ERα was observed. In situ hybridization confirmed the expression of IGF-1R mRNA in the SNpc and immunohistochemistry revealed that almost all TH-ir SNpc DA neurons were IGF-1R-ir (98.5%). Moreover, 33% of glial fibrillary acidic protein (GFAP-ir) cells in the SN were ERβ-ir and 67% of GFAP-ir cells displayed IGF-1R-ir. Thus, the localization of ERβ and IGF-1R on SNpc DA neurons and astrocytes suggests a modulatory role of E2 on IGF-1R and this modulation may impact neuroprotection.
The ovarian steroid hormone estrogen (E2) can act on the central nervous system throughout an organism’s lifespan. The lipophilic nature of steroids allows them widespread access to cells both outside and inside the blood-brain-barrier, coordinating the functions of multiple tissues. Although, the most known effects of E2 are on the reproductive system and on parts of the brain regulating reproduction, recently, it has also become evident that E2 exerts a powerful protective role in the CNS (Chowen et al., 2000; Garcia-Segura et al., 2001; Stein, 2001; Wise et al., 2001; McEwen, 2002). Many studies have focused on E2 protection in brain injuries such as stroke (Barrett-Connor and Bush, 1991; Zhang et al., 1998) and traumatic brain injury (Roof and Hall, 2000a; Roof and Hall, 2000b; Wagner et al., 2004), but there are also suggestive reports that E2 may protect against neurodegeneration in diseases such as Alzheimer’s (Sherwin, 1997) and Parkinson’s disease (Kuppers et al., 2000). The mechanism(s) of E2 neuroprotection has been difficult to elucidate. An important reason for this is that a plethora of E2 effects appears to mediate neuroprotection. For instance, E2 can act via classical genomic mechanisms (Hall et al., 2001) and membrane associated receptors that alter G-protein signaling which activate anti-apoptotic signaling pathways (for review see (Kelly et al., 2003)). In addition, E2 interacts with trophic factors, such as insulin-like growth factor-1 (IGF-1) to prevent neurodegeneration (for review see (Mendez et al., 2005)).
In previous studies using unilateral medial forebrain bundle 6-hydroxydopamine (6-OHDA) lesions as model of Parkinson’s disease in rats, we demonstrated that E2 reduced the loss of substantia nigra par compacta (SNpc) dopamine (DA) neurons and prevented motor deficits produced by 6-OHDA lesions (Quesada and Micevych, 2004). Interestingly, IGF-1 administered centrally or peripherally displays the same effect as E2. Most importantly, in vivo blockaded of IGF-1 receptors (IGF-1R) completely abolished E2 and IGF-1 neuroprotection of SNpc DA neurons. In the 6-OHDA lesion models, these data indicate that E2 protects SNpc DA neurons, and the IGF-1 system has a central role in the mechanism through which E2 exerts its neuroprotective effect on the nigrostriatal DA system. However, an underpinning question is which cell populations, are targeted by E2 and IGF-1.
Within the central nervous system, there is an overlapping expression of estrogen receptor-α (ERα) and estrogen receptor-β (ERβ) mRNA where in some cases the two receptor subtypes are expressed in the same cell (Shughrue et al., 1997). ERα appears to be the dominant receptor mediating E2 actions in many neuronal circuits (McEwen, 2002; Micevych et al., 2003). In the mesencephalon region where nigral DA neurons are located, the ERβ mRNA is the predominant receptor subtype expressed (Shughrue et al., 1997). Immunohistochemical studies confirmed the presence of ERβ-ir in the substantia nigra pars compacta (SNpc) (Shughrue and Merchenthaler, 2001). Although, neurons were thought to be the predominant cell-type expressing estrogen receptor (ER), recent studies in our laboratory have demonstrated that astrocytes also express functional ERs (Sinchak et al., 2003; Chaban et al., 2004).
Neurons and glia cells not only express ERs but are also immunopositive for IGF-1 and IGF-1R (Garcia-Segura et al., 1991; Cardona-Gomez et al., 2000). In addition, E2 can regulate the expression of IGF-1 and IGF-1R in neural tissue (Cardona-Gomez et al., 2001; Cheng et al., 2001; El-Bakri et al., 2004). As previously demonstrated, E2-induced neuroprotective effects on the nigrostriatal DA system are dependent on the activation of the IGF-1R (Quesada and Micevych, 2004). Therefore, besides being extremely important in female reproductive function (Quesada and Etgen, 2001; Quesada and Etgen, 2002), E2/IGF-1 interactions on neurons and glial cells seem also to be critical for neuroprotection during brain injury (Garcia-Segura et al., 2000; Mendez et al., 2005).
To elucidate the anatomical support for the proposed neuroprotective actions of E2 and IGF-1 on SNpc DA neurons we examined whether ERα, ERβ and IGF-1R are expressed on SNpc DA neurons and/or glial cells. Confocal immunofluorescence microscopy was employed to localize ERβ-immunoreactive (ir) and IGF-1R-ir with tyrosine hydroxylase immunoreactive (TH-ir) DA neurons in the SNpc. Moreover, retrograde fluorogold labeled SNpc DA neurons were investigated by examining double-immunofluorescence for ERβ and IGF-1R to verify whether these SNpc DA neurons project to the striatum. We have also compared the patterns of IGF-1R-ir with the expression of IGF-1R mRNA in the SN. Finally, due to the fact that glial cells play a critical role in IGF-1 synthesis (Ye et al., 2004), we examined whether ERβ-ir and IGF-1R-ir were localized in glial fibrillary acidic protein (GFAP-ir) glial cells of the SN.
Adult female (225–250 grams) Long-Evans rats bilaterally ovariectomized (OVX) by the provider, were purchased from (Charles River, Wilmington, MA). OVX animals were used in order to eliminate differences over the cycle and avoid possible differences in ERβ mRNA expression and translation by E2 treatment (Shughrue and Merchenthaler, personal communication). At arrival at UCLA, all animals were housed two per cage in a partially reversed, 12/12 hr light/dark cycle (lights on at 12 midnight) and provided with food and water ad libitum. All of the experimental procedures were approved by the Chancellor’s Animal Research Committee at the University of California, Los Angeles.
rats were deeply anesthetized with sodium pentobarbital (100 mg/kg) and transcardially perfused with physiological saline (4°C) followed by 4% paraformaldehyde in 0.1M Sorensen’s phosphate buffer (4°C). Brains were removed, postfixed for 4–24 hrs in 4% paraformaldehyde solution and then transferred to 15% sucrose in 0.1 M phosphate buffer (pH 7.5), for cryoprotection. Coronal 30 µm sections through the ventral midbrain from four ovariectomized rats were processed simultaneously (Paxinos et al., 1985). Every fourth section was taken through the substantia nigra in a systematic-random manner. Free-floating sections of the ventral midbrain where collected in 0.01 M phosphate buffered saline (PBS, pH 7.4). To quench endogenous peroxidase, all sections were incubated with 10% methanol containing 1% H2O2 for 30 min.
rats were deeply anesthetized with sodium pentobarbital (100 mg/kg) and killed by decapitation. Brains were removed, immediately frozen in 2-methylbutane (Sigma) at a temperature of −50 to −60°C and stored at −80°C. Coronal frozen sections of the brain were serially cut at 20 µm on a cryostat, mounted on Superfrost/plus slides and stored at −70°C until processing. Slides containing fresh-frozen sections were thawed for 10 min, fixed in 4% paraformaldehyde in PBS at 4 °C for 1 hr, washed in PBS (10 mM pH 7.4) 3× for 10 min, penetrated with 0.4% Triton X-100 for 10 min and acetylated for 10 min in 100 mM triethanolamine pH 8.0 containing 0.25% (v/v) of acetic acid. The slides were rinsed in PBS (3 × for 8 min) and distilled water, dehydrated in 50% and 70% ethanol and air-dried.
Tyrosine hydroxylase (TH) is an affinity purified sheep polyclonal antibody (Chemicon International, Temecula, CA). Isulin-like growth factor-1 receptor (IGF-1R), a mouse monoclonal antibody (Chemicon Research Products, San Diego, CA) was used. For ERβ immunocytochemistry; Z8P, an affinity purified rabbit polyclonal antiserum raised against amino acid 468–485 of the C-terminal domain of the mouse ERβ (Zymed laboratories, San Franciso, CA). The Z8P antibody was wa previously characterized (Shughrue and Merchenthaler, 2001) and ERβ-ir is present in cells that express ERβ mRNA throughout the brain of female rats including the SNpc. Another ERβ antibody used was, Ab-1, a rabbit polyclonal antiserum raised against amino acids 467–485 of the C-terminal domain of the Rat ERβ (Oncogene Research Products, San Diego, CA), a region with no homology to ERα, and recognizes ERβ of rat origin by Western blotting (Li et al., 1997). For ERα, a rabbit polyclonal antibody raised against a peptide corresponding to amino acid 580–599 of the C-terminal domain of the mouse ERα (MC-20, Santa Cruz Biotechnology, Santa Cruz, CA). Glial fibrillary acidic protein (GFAP) immunoreactivity was detected using a rabbit polyclonal antibody raised against human GFAP (G9269; Sigma, St. Louis, MO).
For stereologic analysis of ERβ-ir, we employed the Z8P antibody, because it predominately labeled the nucleus of TH-ir SNpc DA neurons. To quantify the number of ERβ-ir, IGF-1R-ir in TH-ir SNpc DA neurons, sections were incubated in blocking solution (5% normal donkey serum (NDS), 0.3% Triton X-100 in PBS) for 1 hr in order to block non-specific binding. Then, sections were treated with 0.2% Triton X-100 (3 ×) for 10 minutes followed by 0.1 M glycine for 30 min. Sections were incubated with anti-TH (1:6000; Chemicon, Temecula, CA) overnight at 4°C, followed by biotinylated anti-sheep IgG and avidin-biotin-peroxidase complex (Vector Laboratories) for 1 hr at RT. TH immunoreactivity was visualized by 3,3-diaminobenzidine tetrahydrochloride (DAB) (Sigma). Sections were incubated for 48–72 hr at 4 °C with Z8P (2.5 µg/ml in 1% NDS, 0.3% Triton X-100 in PBS) or anti-IGF-1R (1:600 in 1% NDS, 0.3% Triton X-100 in PBS). After incubation with primary antibody, sections were incubated with biotinylated anti-rabbit IgG or biotinylated anti-mouse IgG (Vector Laboratories) for ERβ and IGF-1R, respectively for 1h at RT. Sections were then incubated in avidin-biotin-peroxidase complex (Vector Laboratories) for 1 hr. After series of washes, ERβ and/or IGF-1R were visualized by DAB-nickel peroxidase reaction (Vector DAB substrate kit Cat # SK-4100, Vector Laboratories). Sections were mounted on Superfrost Plus slides, dehydrated through a series of graded alcohols cleared in xylene, and coverslipped with Permount mounting medium (Fisher Scientific).
Total numbers of ERβ and/or IGF-1R that are expressed in TH immunopositive neurons of the SNpc were estimated using the optical fractionator method (West et al., 1991) with stereoinvestigator® software (MicroBright-Field, Colchester, VT). The precision of the serial analysis was assessed by the coefficient of error (p < 0.1) (West and Gundersen, 1990; West et al., 1996). Briefly, each subfield of the SNpc was manually outlined at low magnification (5× objective) and a grid of 150 × 150 µm was superimposed by the software. The SNpc was delimited rostrally by the optic tract, caudally by the level of decussation of the superior cerebellar peduncle, ventrally by the pars reticulate subdivision of substantia nigra, and medially by the third cranial nerve rootlets. A 63× oil immersion objective (1.4 NA) was used to achieve optimal optical sectioning during dissector analysis. Quantitative analysis was performed by placing an unbiased optical dissector frame (50 × 50 µm = 2,500 µm2 in the x-y-axis) over the contour outlining the SNpc. The final post-processing thickness of the sections was measured by the microcator, resulting in a 15 µm averaged mounted section thickness. A 1.5 – 2 µm guard zone was used to exclude artifacts at tissue surface. For colocalization quantitative analysis, only those TH-ir cells that clearly demonstrated nuclear ERβ were counted.
To enhance the sensitivity of the immunocytochemical reaction the tyramide signal amplification method (TSA-Direct, Perkin Elmer Life and Analytical Sciences, Boston, MA) was employed. Sections were incubated 48–72 hr at 4°C with anti-TH (1/500 in 1% NDS, 0.3% Triton X-100 in PBS); anti-IGF-1R (1:600 in 1% NDS, 0.3% Triton X-100 in PBS); Z8P (1/400 in 1% NDS, 0.3% Triton X-100 in PBS) or Ab-1 (1/200 in 1% NDS, 0.3% Triton X-100 in PBS). After incubation with primary antibodies, sections were washed in PBS (3 ×) for 10 min and incubated for 1 hr at RT with the secondary antibodies. To visualize TH-ir, rodamine (RITC)-conjugated Donkey Anti-Sheep (1:100; Jackson ImmunoResearch lab, INC; Baltimore, PA) secondary antibody was used. To visualize ERβ-ir, sections were incubated with biotinylated donkey anti-rabbit serum (1:1000; Jackson ImmunoResearch Lab, West Grove, PA). These sections were then incubated with HRP-conjugated streptavidin (1:100; TSA kit in TNB buffer) for 30 min at RT, rinsed in PBS (3 ×) for 5 min, and then incubated with fluoroscein conjugated tyramide (FITC-conjugated tyramide; 1:100 in 1 × amplification diluent; TSA kit) for 10 min at RT. To visualize IGF-1R, sections were incubated with biotinylated donkey anti-mouse serum (Vector Labs; 1:200 in PBS) for 1 hr. Sections were then washed in PBS (3 ×) for 10 min and incubated with HRP-conjugated streptavidin (1:100; TSA kit in TNB Buffer) for 30 min at RT, rinsed in PBS (3 ×) for 5 min, incubated with fluoroscein conjugated tyramide (FITC-conjugated tyramide; 1:100 in 1 × amplification diluent; TSA kit) for 10 min at RT. Tissue sections were washed in PBS (3 ×) for 10 min, mounted on Superfrost Plus slides (Fisher Scientific, Pittsburgh, PA), air-dried and coverslipped using Vectashield mounting medium (Vector Laboratories, Burlingame, CA).
For double immunofluorescence for ERα/β or IGF-1R in TH-ir nigral DA neurons, sections were incubated for 48–72 hr at 4°C with a cocktail of antisera raised in different species containing anti-TH with either IGF-1R or Ab-1 or Z8P. Double immunofluorescences for ERβ or IGF-1R in GFAP-ir glial cells, sections were processed as mentioned above. Following amplification of either IGF-1R and/or ERβ with the TSA kit as previously described above, sections were incubated in blocking solution (5% NGS in PBS for 1 hr) and incubated for 1 hr at 4 °C with rabbit anti-GFAP (G9269; Sigma, St. Louis, MO; diluted 1:50) in 1% NGS and 0.3% Triton X-100 in PBS. GFAP-ir was detected by incubating the sections for 1 hr in Texas Red –conjugated goat anti-rabbit antibody (Molecular Probes; diluted 1:1000; 1% NGS). Colocalization for ERβ or IGF-1R in GFAP-ir cells in the SN was analyzed by the (Carl Zeiss Inc, Thornwood NY) LSM 3.2 software. Z-stacks of images were taken using a 40× oil immersion objective. An area of 1000 mm2 was delineated from each Z-stack. Quantitative data is expressed as an area of a colocalized number of pixels.
The following controls were performed: omitting the primary or secondary antibodies, dilution series and nonimmune rabbit or mouse immunoglobulin G (IgG; 50 µg/µl). The following peptides were used for preabsorbance: the epitope peptide to anti-ERα (MC20, Santa Cruz Biotechnology); and peptide to block the anti-ERβ Z8P antibody and Ab-1 antibody was preabsorbed with the 19 –amino acid sequence of ERβ (amino acids 467–485)(PA1-310B; Affinity Bioreagents). All staining was blocked by preabsorption. The pattern of staining for each antibody was the same in double or single labeled sections. For double labeling, the order of primary antibodies was reversed and homologous and heterologous absorption controls were done. Nonspecific immunoflourescent staining, cross-immunostaining, or bleeding-through were not observed.
Immunofluorescence double-labeled profiles were scanned with a Zeiss Axioskop LSM 510 laser scanning confocal microscope system equipped with an Axiocam CCD camera and a digital image analysis system (Zeiss USA Inc.). The excitation source was a krypton-argon laser (Coherent, Santa Clara, CA) with output at 488,568 and 633 nm. FITC was visualized with a 488 nm excitation and a 515–540 nm bandpass emission filter. Rhodamine was visualized with a 530 nm excitation and 560–610 nm bandpass emission filter. Z-stacks were obtained by using identical settings for laser intensity, confocal aperture, scan speed and Z-stack step size (0.5 – 1.0 µm). All final images were adjusted for contrast and background using Adobe Photoshop 5.0 (Mountain View, CA).
OVX rats were implanted with a guide cannula aimed at the striatum (coordinates: antero-posterior, bregma +0.48 mm; lateral, ± 3.0 mm; and doso-ventral, −5.0 mm, (Swanson, 1992). A 29 gauge cannula containing the retrograde tracer Fluorogold™ (FG; 5% w/v dissolved in 0.9% saline; Fluorochrome, Denver, CO) was used to slowly eject 1 µl of FG over 5 min (Iravani et al., 2002). The cannula was connected to a 25 µl Hamilton syringe via a plastic tube. The cannula was allowed to remain in place an additional 5 min to allow diffusion of drug away from the cannula tip. Seventh day after FG injection, animals were perfused.
Immunocytochemical labeled sections from FG injected animals were visualized with a Axioflouroskop 2 equipped with epiflourescent illumination and Axiovision 3.1 software. Digital pictures were taken with an Axiocam CCD camera mounted on the microscope (Zeiss Inc. USA). To detect the immunolabeling of either ERβ-ir and/or IGF-1R-ir in striatal projecting SNpc DA neurons, SN sections from rats injected with FG into the striatum were processed for immunofluorescence for either ERβ and/or IGF-1R immunohistochemistry as described above. The number of FG SNpc cells immunolabeled for either ERβ and/or IGF-1R were averaged at 4 different levels corresponding to −4.8 mm, −5.1 mm, −5.4 mm and −5.7 mm with respect to bregma (Paxinos et al., 1985) and expressed as a percentage. FG was visualized with a 333 nm excitation and wideband ultraviolet emission filter at 408 nm. FITC was visualized with a 488 nm excitation and a 515–540 nm bandpass emission filter.
The IGF-1R probe was complementary to 15 bases of 5’ untranslated region encoding the signal peptide, which correspond to the first 53 amino acids of the alpha subunit (Werner et al., 1989). The cDNA EcoR I/Sma I fragment (265 bp) was subcloned into a pGem-3Z transcription vector (gift from Dr. D. LeRoith; National Institutes of Health, Bethesda, MD, USA), and linearized with EcoRI (antisense) or BamHI (sense) and isotopically labeled riboprobes were generated by in vitro transcription using SP6 (antisense) and T7 (sense) polymerases in the presence of 35S-UTP and 35S-CTP (Amersham Biosciences Corp. Piscataway, NJ). Riboprobes were purified through NucTrap purification columns (Stratagene), diluted in hybridization buffer (100 mM Tris pH 7.5, 600 mM NaCl, 1 mM EDTA, 0.1 mg/ml sonicated salmon sperm, 0.5 mg/ml t-RNA, 1× Denhardt’s, 10% dextrane sulfate, 50% formamide) to 100,000 cpm/µl and used within 48 hr.
Tissue sections were covered with hybridization buffer (80 µl) containing antisense or the sense riboprobes, coverslipped to prevent evaporation and hybridized for 16–18 hr at 58°C in a moist chamber containing 50% v/v formamide. Following hybridization, coverslips were carefully removed, sections were washed with 2 × sodium citrate (SSC) and 1 × SSC for 15 min each at room temperature and exposed to RNAse buffer (20 µg/ml RNAse A, 1 U/ml RNAse T1, 10 mM Tris-HCl, 0.5 M NaCl, 1 mM EDTA) at 37°C for 40 min. Afterwards, slides were washed in decreasing concentrations of SSC to a final stringency of 0.1 × SSC at 62 °C for 1 hr. Finally slides were washed in double distilled water, dehydrated through graded ethanols and air dried. Slides were dipped in K5 photoemulsion (Ilford, Paramus, NJ), diluted 1:1 with double-distilled water. Emulsion coated slides were air-dried in the dark for 12 hr, packed into light-tight boxes and exposed at 4°C for 21–28 days, developed (D-19 developer) and fixed (Kodak fixer) both products from (Kodak Co, Rochester, NY). The slides were counterstained with Toluidine blue, dehydrated and coverslipped with DPX (BDH Chemical Ltd. London, UK).
The distribution of ERβ in the SNpc was examined using two distinct antisera (Z8P and Ab-1). Single immunolabeling with these ERβ antibodies displayed similar immunostaining patterns for ERβ, especially in the SNpc (Fig 1E and 1F). Throughout the mesencephalon, the ventral tegmental area (VTA) had the densest number of ERβ-ir cells, which expanded laterally to the SNpc, substantia nigra pars reticularis and substantia nigra pars lateralis. Within the SN, both ERβ antibodies labeled a high density of cells in the SNpc compared to the pars reticularis or pars lateralis (Fig 1E and 1F). Moreover, scattered ERβ-ir were also noted in the interpeduncular nucleus, periaqueductal gray, mesenphalic reticular nucleus, medial lamniscus, raphe nuclei and optic tectum as previous confirmed by (Shughrue and Merchenthaler, 2001)(data not shown). For ERα-ir, the rabbit polyclonal antibody to ERα did labeled an abundant amount of ERα in the mediobasal hypothalamus, a region abundant with ERα; however, only a scattered number of ERα-ir cells were detected in the VTA, but none were observed in the SN (data not shown).
IGF-1R mRNA and protein were demonstrated by in situ hybridization and immunocytochemistry in the mesencephalon region of OVX rat (Fig 1C and and2B).2B). Within the ventral midbrain, abundant hybridization signal for IGF-1R was observed in the VTA and SN (Fig 2B). Within the SN, the SNpc had the highest density of IGF-1R-ir cells followed by the substantia nigra pars reticularis (Fig 1C and Table 1). Other regions that had strong IGF-1R mRNA expression and protein immunoreactivity, include the interpeduncular nucleus, and the superior and inferior colliculus, with moderate expression in the medial geniculate (data not shown).
The present double immunolabeling studies revealed that in the midbrain, the VTA had the majority of TH-ir cells that express ERβ, followed by the SNpc with a few scattered double-labeled cells in the SN pars reticulata and pars lateralis (Table 1). However, the cellular localization was significantly different between Ab-1 and Z8P ERβ antibodies. The Ab-1 ERβ antibody showed ERβ-ir predominantly in the cytoplasm of TH-ir SNpc DA neurons with sparse nuclear labeling (Fig 3A), whereas the Z8P ERβ antibody predominately labeled the nucleus with low extra-nuclear staining in TH-ir DA SNpc neurons (Fig 3B). Unbiased stereologic analysis revealed that in the SNpc, 39.8% ± 4.7 of TH-ir cells were double-labeled with ERβ-ir (Figure 4A & 4B; Table 2). A majority of double-labeled neuronal somata were oval in shape, with an average neuronal diameter of 14.8 ± 0.7 µm and nuclear diameter of 8.3 ± 0.4 µm. Confocal microscopy revealed a significant colocalization of ERβ- and TH-ir in SNpc DA neurons (Fig 5A–C).
Virtually all of TH-ir cells were immunopositive for IGF-1R within the mesencephalon. The VTA demonstrated an abundant colocalization (Table 1). Within the SN, the SNpc demonstrated a complete colocalization 98.5% ±0.5, which expanded laterally to the SN pars lateralis and reticularis (Table 1; Fig 4C & 4D; Table 2). Confocal microscopy revealed that IGF-1R-ir was predominantly localized to the membrane in TH-ir cells although some cytoplasmic labeling was also observed (Fig 6A–C).
To verify that TH-ir DA neurons in the SNpc that are ERβ-ir and/or IGF-1R-ir project to the striatum, FG was injected into the striatum to retrogradely label SNpc DA neurons (Fig 1B). 43 ± 8.9 % of these FG striatal projecting neurons were ERβ-ir (Fig 5D–F). Likewise, the vast majority of FG labeled neurons were immunopositive for IGF-1R (97.8 ± 9.8 %) (Fig 6D–F).
The morphology of GFAP-ir glial cells in the SN of OVX rats were stellate, hypertrophic, thorn-shape and coiled (Fig 7A and 7D)(Lu et al., 2003). Analysis of colocalization revealed that ERβ-ir with Z8P antibody labeled 33 ± 8.9 % GFAP-ir astrocytes. ERβ-ir was distributed in the nucleus, somal cytoplasm and in the processes of GFAP-ir cells (Fig 7C). 67 ± 10.5 % of the GFAP-ir cells contained IGF-1R-ir. The IGF-1R was primarily localized in the perikarya and processes of GFAP-ir astrocytes (Fig 7F).
In these studies we demonstrate that SNpc DA neurons and GFAP-ir astrocytes in the SN express ERβ and IGF-1R in OVX rats. These results provide anatomical basis for the proposed mediated neuroprotection of E2 and IGF-1 in SNpc DA neurons. The present immunohistochemical study demonstrates the localization of ERβ in SNpc DA neurons without a parallel localization of ERα in the SN and supporting the role for ERβ in neuronal survival of SNpc DA neurons (Sawada et al., 2000; Wang et al., 2001). In addition, since both glial cells and neurons were immunoreactive for ERβ and IGF-1R, suggest a complex interaction in which paracrine and/or autocrine functional regulation of SNpc DA neurons.
E2 and IGF-1 interactions have been shown to be important in physiological regulation of CNS activity such as regulation of growth, sexual maturation and adult neuroendocrine function (Fernandez-Galaz et al., 1997; Miller and Gore, 2001; Quesada and Etgen, 2001; Quesada and Etgen, 2002). Furthermore, E2 and IGF-1 interactions have also implemented in protecting the CNS in different experimental models of neurodegenerative diseases (Azcoitia et al., 1999b; Garcia-Segura et al., 2001; Carro et al., 2003). For instance, our previous studies demonstrated that E2 acting through activation of IGF-1R produced an 80% survival of SNpc DA neurons treated in vivo with 6-OHDA, compared to 40% survival after 6-OHDA lesion in the midbrain (Quesada and Micevych, 2004). In the present study, stereologic analysis reveals that 39.8% of SNpc DA neurons have ERβ-ir, however 33% of GFAP-ir glial cells in the SN also displayed ERβ-ir. E2 has been shown to increase the expression of growth factors, such as IGF-1 in glial cells (Fernandez-Galaz et al., 1997; Zhang et al., 2004) suggesting that astrocytic IGF-1 may act to spare DA neurons from neurodegenerative processes. Through these actions, E2 may protect the remaining non-expressing ERβ SNpc DA neurons indirectly by the release of IGF-1 from astrocytes. A hypothetical model for these interactions is presented in Figure 8. The localization of ERβ on GFAP-ir cells in the SN provide cellular support for E2 actions on these glial cells to express and secrete beneficial anti-neurotoxic signals to SNpc DA neurons. Therefore, ERβ and IGF-1R may be relevant for the coordinated actions of E2 and IGF-1 on both SNpc DA neurons and glial cells.
Our experiments employed in situ hybridization and immunocytochemistry to reveal the presence of IGF-1R in the SNpc of OVX female rats. Almost all of SNpc DA neurons (98.5%) and a vast majority of GFAP-ir glial cells in the SN are IGF-1R-ir (67.0%). These results are consistant with the idea of IGF-1 playing a major neuroprotective action on the nigrostriatal DA system, which accounts for the protection of SNpc DA neurons against 6-OHDA lesions, when animals receive centrally and peripherally administered IGF-1 (Quesada and Micevych, 2004). Therefore, if the IGF-1 system is activated by neuronal injury, then antagonism of IGF-1R would exacerbate the neuronal injury as we and others previously shown (Quesada and Micevych, 2004).
Although there is some controversy about the distribution of ERβ-ir in the SN (Creutz and Kritzer, 2002) our data support the presence of ERβ-ir, in agreement with others (Shughrue and Merchenthaler, 2001; Zhang et al., 2002; Mitra et al., 2003). In contrast, a recent report using the Z8P antibody failed to detect ERβ-ir in the SNpc or in GFAP-ir cells of the SN (Creutz and Kritzer, 2002). These authors found colocalization of ERβ-ir in TH-ir cells located in a region immediately dorsal to the SNpc, but not below. To examine this possibility, the present study-combined tract tracing with ERβ immunohistochemistry and clearly demonstrates ERβ-ir in SNpc DA neurons that project to the striatum, confirming the previous report (Shughrue and Merchenthaler, 2001).
The cellular distribution of ERβ with the Z8P antibody in SNpc DA neurons and glial cells is in agreement with previous reports demonstrating nuclear ERβ-ir and cytoplasmic labeling in GFAP-ir astrocytes in the interpedunuclar nucleus of female rats (Zsarnovszky et al., 2002) and in cultured cortical astrocytes (Chaban et al., 2004). In addition, the nuclear and cytoplasmic distribution of ERβ in SNpc DA neurons and in glial cells throughout the brain has been observed with the Ab-1 antibody (Cardona-Gomez et al., 2000; Ravizza et al., 2002). Our present results are in agreement with others demonstrating differences in cellular localization of ERβ by Z8P (Shughrue and Merchenthaler, 2001; Blurton-Jones and Tuszynski, 2002; Ravizza et al., 2002; Chakraborty et al., 2003) and Ab-1 (Azcoitia et al., 1999a; Ravizza et al., 2002). As discussed by Shughrue and Merchenthaler (2001) the difference in staining between these two antibodies (Z8P and Ab-1) may be a difference between rat vs. mouse peptide (4aa), peptide conjugation, inoculation, and purification resulting in antisera that are vastly different in both their sensitivity and specificity for rodent ERβ. Moreover, using different ERβ antibodies than in the present study, cytoplasmic staining on SNpc DA neurons has been demonstrated (Ravizza et al., 2002; Zhang et al., 2002; Mitra et al., 2003). Recently, there has been an increasing body of evidence suggesting that membrane-bound ERs are present in the brain (Nishio et al., 2004) and mediate neuroprotection as demonstrated in cultured mesencephalic DA neurons (Bains, 2005) and cultured cortical astrocytes (Bains, 2005; Dhandapani et al., 2005). Despite the subcellular distribution of ERβ-ir in the SNpc, the present results suggest that the neuroprotective effects of E2 are mediated by ERβ which may act through transcriptional regulation or rapid signaling cascades which can lead to signaling transcription regulation (Dubal et al., 1998; Garcia-Segura et al., 2001; Wise et al., 2001; Ivanova et al., 2002; Boulware et al., 2005).
E2 and IGF-1 can mutually regulate each others receptor expression (Westley and May, 1994; Cardona-Gomez et al., 2001). For example, IGF-1 can activate ER, in the absence of E2, suggesting that ER may play an important regulating role in IGF-1R signaling (Ma et al., 1994; Cardona-Gomez et al., 2001). The localization of ERβ- and IGF-1R-ir on nigral DA neurons and GFAP-ir glial cells in the SN may provide the anatomical support for these interactions between E2 and IGF-1 (Mendez et al., 2003). In this context, E2 may upregulate the expression of IGF-1/IGF-1R or ERβ may be important for IGF-1 neuroprotection by which IGF-1 regulates the expression of ERβ in SNpc DA neurons. Alternatively, MAPK and PI3K, two rapid intracellular kinases activated by E2 and IGF-1 have been shown to protect neurons against toxic effects in vitro (Singer et al., 1999; Singh et al., 2000). Therefore, coexpression of ERβ and IGF-1R in the same cell may allow a cross-talk between ERβ and IGF-1R signaling pathways mediating neuroprotection (Mendez et al., 2005). Indeed, future experiments are needed to examine the intracellular signaling pathways used by E2 and IGF-1 in the neuroprotection of nigral DA neurons.
In addition to genetic and environmental factors that contribute to vulnerability for neurodegeneration, age is one of the most important risk-factor. As women age and become menopausal, systemic E2 levels drop to undetectable levels and IGF-1 levels are significantly reduced (Lamberts et al., 1997), suggesting that the interaction of E2 and IGF-1 signaling pathways in neuroprotection may be relevant concerning the increased neurodegenerative incidence at old age. Moreover, in human brains, the SN has an abundant expression of IGF-1R (De Keyser et al., 1990). These results along with past studies suggest that decreased IGF-1R expression levels and/or IGF-1R signaling capacity may be underlie susceptibility for deficient in neurodegenerative diseases such as Parkinson’s disease. Within the context of this working hypothesis, the survival of SNpc DA neurons is dependent in a trophic influence of E2 and IGF-1. Therefore, the IGF regulatory system will be a prime target for future studies into the pathogenesis of several age and sex hormone related degenerative diseases providing important clues as to how E2 is protective.
Grant Sponsor: National Institute of Health DE016063-01 and the American Parkinson’s Disease Association NSO 39495