|Home | About | Journals | Submit | Contact Us | Français|
Many human immunodeficiency virus (HIV) proteins including Tat are produced by HIV-infected astrocytes and secreted into the brain resulting in extensive neuronal damage that contributes to the pathogenesis of HIV dementia. The neuroprotective hormone 17β-estradiol (E2) is known to negatively regulate the HIV transcriptional promoter in human fetal astrocytes (SVGA cell line) in a Tat-dependent manner. In the present study we extended our investigation in HIV-infected SVGA cells and found a reduction in HIV p24 levels following E2 treatment in comparison to control. Although many E2-mediated events occur through estrogen receptor alpha (ERα), we found low levels of ERα mRNA and failed to detect ERα protein in SVGA cells. Paradoxically, when ERα was overexpressed the E2-mediated decrease in Tat transactivation of the promotor was prevented. To determine whether ERα expression is altered in the human brain following HIV infection, postmortum hippocampal tissue was obtained from cognitively normal HIV− and HIV+ patients, patients diagnosed with either mild cognitive/motor disorder (MCMD) or HIV-associated dementia (HAD). Immunohistochemistry and quantitative real-time PCR (qRT-PCR) for ERα and glial fibrillary acidic protein (GFAP) showed that ERα mRNA levels were not significantly different between groups, while GFAP increased in the hippocampus in the HIV+ compared to the HIV− group and was decreased in the MCMD and HAD subgroups compared to HIV+ controls. Notably the ratio of ERα-positive reactive astrocytes to total reactive astrocytes increased and significantly correlated with the severity of cognitive impairment following HIV infection. The data suggest that E2 would have the most dramatic effect in reducing HIV transcription early in the disease process when the subpopulation of astrocytes expressing ERα is low.
HIV-associated dementia (HAD) develops in approximately 30% of all HIV-infected patients, while more than 60% of all HIV-infected patients develop neurological impairment.1 Astrocytes are important participants in the neuropathogenesis of HAD and unlike microglia, harbor a nonproductive infection whereby viral proteins are expressed but replication of the viral genome does not occur.2 One particular viral protein produced by astrocytes is Tat. Tat is essential for viral replication and activates the transcription of all viral proteins3–5 by binding the transactivation response region (TAR) within the 5′ terminus of all retroviral mRNAs6,7 and stimulating the elongation efficiency of RNA polymerase II.8,9 In addition, Tat is secreted by infected astrocytes into the brain and causes extensive neuronal damage10 directly or indirectly by stimulating the release of neurotoxic factors from infected astrocytes.11,12 Given the many ways in which Tat and astrocytes together participate in the pathogenesis of HAD, effective neuroprotective strategies are needed that either target inflammation and neuronal damage directly or indirectly by inhibiting the production of toxic viral proteins.
17β-Estradiol (E2) is one factor that has been shown to be neuroprotective and antiinflammatory.13,14 E2 protects against Tat-induced neurotoxicity,15–17 oxidative stress-induced damage, microglial activation, and concomitant secretion of inflammatory mediators13,14,18 and alters neuronal morphology in regions of the brain involved in cognition.19,20 Furthermore, our laboratory previously reported that E2 attenuated Tat-induced HIV long terminal repeat (LTR) promoter activation in SVGA cells.21 Because estrogen receptors are expressed by neurons and glia, estrogen may mediate neuroprotection through receptor-dependent processes. Indeed the protective effect of estrogen against ischemia-induced damage is dependent on the presence of estrogen receptor alpha (ERα).22,23 Furthermore, astrocytes have been implicated in the neuroprotective effects of estrogen in the brain,24–26 The fact that astrocyte expression of ERα is upregulated following injury to the rat,24 primate,27 and human brain28 suggests a role for astrocyte-derived ERα in estrogen-mediated neuroprotection. The present study was designed to investigate whether the E2-mediated suppression of Tat transactivation in astrocytes was dependent on the activation of estrogen receptors and, furthermore, to test whether ERα is upregulated in astrocytes in the HIV brain.
An immortalized astrocyte cell line, SVGA,29 stably transfected with the HIV-LTR promoter driving chloramphenicol acetyltransferase (CAT) (SVGA-CAT) was a kind gift from A. Nath, Johns Hopkins University. HEK293 cells were a kind gift from J. Satin, University of Kentucky, HeLa cells were a kind gift from A. Nath and MCF-7 cells were a kind gift from M. Kilgore. Cells were grown in DMEM+10% charcoal stripped fetal bovine serum (FBS)+1% penicillin/streptomycin and maintained at 37°C in 5% CO2. E2 solutions were prepared in ethanol and diluted so that the final ethanol concentration was 0.01%. Cells were seeded into 24-well plates and treated with E2 or vehicle. After 48h cells were transfected with 0.8μg Tat1-72 or pcDNA (empty vector control) using Lipofectamine 2000 (Invitrogen, Carlsbad, CA). After 24–48h cells were harvested and assayed for CAT levels by ELISA (Roche Applied Science, Indianapolis, IN) as per the manufacturer's instructions. To determine the effect of estrogen on HIV transcription SVGA cells were pretreated with 1nM doses of E2 or vehicle for 48h and then infected for 2h with 2000pg/ml p24 collected from the supernatant of H9/HTLV-IIIB (NIH 1983) cells.30 SVGA infection was monitored by the measurement of p24 protein by ELISA (Perkin-Elmer, Waltham, MA) as per the manufacturer's instructions.
A total of 13 autopsy cases (3 from HIV-negative controls and 10 from HIV-infected individuals) were used in the present study. Frozen blocks of postmortem human hippocampal tissue as well as 5-μm-thick paraffin-embedded human hippocampal sections were obtained from the National NeuroAIDS Tissue Consortium (NNTC; Federal grant number N01MH32002). Fresh hippocampal tissue for RNA isolation and formalin-fixed paraffin-embedded hippocampal sections for immunohistochemical analysis were obtained from age-matched male and female subjects with minimal neuropathological diagnosis. Fresh and fixed tissue was obtained from three HIV-negative patients, three HIV-infected patients with normal cognitive function, three HIV-infected patients with mild cognitive or motor disorder (MCMD), and four HIV-infected patients with HAD. Prior approval for the use of human material in our studies was obtained from the University of Kentucky Medical Institutional Review Board.
Total RNA from SVGA, HEK293, or MCF-7 cells was extracted with TRIzol Reagent (Invitrogen, Carlsbad, CA). Total RNA from blocks of frozen human hippocampal tissue weighing between 11mg and 311mg was isolated according to the method published by Chomczynski and Sacchi.31 Briefly, tissue was homogenized in 250μl guanidinium thiocyanate buffer containing 4M guanidinium thiocyanate, 25mM sodium citrate, 0.5% sarcosyl, and 0.7% 2-mercaptoethanol in DEPC water followed by phenol-chloroform extraction. First strand cDNA synthesis was carried out with SuperScript II and random hexamers (Invitrogen, Carlsbad, CA). Reverse transcription was performed at 42°C for 50min followed by 70°C for 15min. cDNA was quantified using the Quant-iT Oligreen ssDNA Assay Kit (Molecular Probes, Invitrogen, Carlsbad, CA). Primer sets used in the real time PCR reactions were as follows: ERα 5′ GTGCCTGGCTAGAGATCCTG and 3′ GATGTGGGAGAGGATGAGGA, GFAP 5′ CTGGAGGTTGAGAGGGACAA and 3′ TCCTCCAGCGACTCAATCTT. Primers for reference gene expression (HPRT, RPL32, and H3.3) were obtained corresponding to published sequences.32,33 Amplification was performed in Mx3000P 96-well plates (Stratagene) in triplicate in a Mx3000P Sequence Detection System (Stratagene, La Jolla, CA) using 1ng cDNA, 125nM forward and reverse primer, 0.75μl Reference Dye (1:500), and 25μl 2× SYBRGreen Brilliant Master Mix (Stratagene, La Jolla, CA) in a 50μl reaction. Thermalcycler conditions were 95°C for 10min followed by 40 cycles of 95°C for 30s, 60°C for 30s, and 72°C for 30s and 1 cycle of 95°C for 1min and 55°C for 30s. MCF-7 breast cancer cells were used as templates for ERα, H3.3, RPL32, and HPRT mRNA standard curves and HIV-negative human hippocampus was used for the generation of GFAP mRNA standard curves. Real time PCR data were normalized to the geometric mean of all reference genes according to the method previously described.34
SVGA cells were plated into 12-well plates and transfected with 1μg ERα using Lipofectamine 2000. After 48h the cells were fixed with 1:1 methanol:acetone, rinsed, blocked in 5% normal donkey serum (Jackson Immuno Research, Westgrove, PA), and incubated overnight at 4°C in rabbit anti-ERα (MC-20) (1:200; Santa Cruz Biotech, Santa Cruz, CA, catalog #sc-542). ERα was visualized with the use of antirabbit Alexa 594-conjugated secondary antibody (1:500) applied for 1h at room temperature (RT). Cells were rinsed and left in phosphate-buffered saline (PBS) while images were captured. Then 5-μm-thick paraffin-embedded human hippocampal sections were processed for ERα and GFAP immunohistochemistry with the use of the Antibody Amplifier (ProHisto, Columbia, SC). Sections were incubated at 60°C for 30min prior to deparaffinization in xylene and ethanol. Sections were hydrated and double-label fluorescence immunohistochemistry was performed.35 Briefly, sections were rinsed in 0.05M Tris-buffered saline (TBS) pH 7.2 (3×10min) and boiled in 0.05M citric acid-buffered saline (CBS) pH 6.0 for 2×10min in a conventional microwave oven. After rinsing, sections were incubated for 60min at RT in 5% normal goat serum (Jackson ImmunoResearch, Westgrove, PA) and incubated at 4°C for 36h in rabbit anti-ERα (1:400 MC-20, Santa Cruz Biotech. Inc., Santa Cruz, CA, catalog #sc-542). Visualization was achieved with the use of antirabbit biotinylated (1:200 applied for 1h at RT) followed by Cy3-conjugated streptavidin (1:500 applied for 45min at RT) secondary antibodies. Sections were rinsed and incubated in 0.05M glycine–HCl buffer saline pH 2.2 for 2h at RT. After rinsing, sections were incubated in monoclonal anti-GFAP (1:800 Sigma, St. Louis, MO, catalog #G3893) for 60min at 37°C followed by 4°C overnight. Visualization was achieved with the use of FITC-conjugated goat antimouse (1:200, Jackson ImmunoResearch, Westgrove, PA). After extensive washing, sections were coverslipped with Vectashield Hardset Mounting Medium with DAPI (Vector Laboratories, Burlingame, CA). Peroxidase staining for ERα and GFAP was performed.36,37 Sections were deparaffinized, boiled in CBS as before, and incubated in 0.3% hydrogen peroxide in TBS containing 0.3% Triton X-100 for 30min. After rinsing in TBS, sections were blocked in 5% normal goat serum for 1h at RT followed by incubation at RT overnight in rabbit anti-ERα (1:75 MC-20, Santa Cruz Biotech. Inc., Santa Cruz, CA, catalog #sc-542) or monoclonal anti-GFAP (1:1000 Sigma, St. Louis, MO, catalog #G3893). Visualization was achieved with antirabbit and antimouse biotinylated secondary antibodies, respectively (1:200 applied for 1h at RT). After several washes in TBS, sections were incubated in avidin-biotin peroxidase complex (Vectastain Elite ABC Kit, Vector Laboratories, Burlingame, CA) in TBS for 45min at RT and developed using 3,3′-diaminobenzidine (DAB) tetrahydrochloride tablets (Sigma, St. Louis, MO).
Sections colabeled for ERα and GFAP were imaged on a Nikon Eclipse TE 2000-E inverted fluorescence microscope (Nikon Instruments, Inc., Melville, NY) with the use of a 10× objective and filters equipped with excitation/emission spectra to detect fluorochromes FITC, Cy3, and DAPI. Images were captured and merged using Spot RT Software v3.5.1 through a Spot RT color camera (Diagnostic Instruments Inc., Sterling Heights, MI). Images were opened in Adobe Photoshop v7 and adjusted for brightness and contrast. Using the zoom tool, GFAP-immunoreactive (IR) cells containing ERα per unit area (0.9mm×0.65mm) were marked and manually counted in each image. Three to five regions within the dentate gyrus were analyzed for each section. Sections processed for peroxidase staining of ERα or GFAP were examined by light microscopy and digitalized through a Spot RT camera.
Nuclear and cytoplasmic fractions were isolated from cell lysates using the NE-PER kit (Pierce, Rockford, IL). Total protein was determined by a microplate Lowry-based protein assay (Bio-Rad). Twenty five micrograms of protein from each sample was separated on a 10% Tris–HCl polyacrylamide precast gel (Bio-Rad). The separated proteins were then transferred to nitrocellulose membranes and blocked in Odyssey blocking reagent for 1h at RT and the following primary antibodies applied: rabbit anti-ERα (1:200, Santa Cruz Biotech., Santa Cruz, CA, catalog #sc-542) and monoclonal antiactin (1:500; Sigma, St. Louis, MO, catalog #A4700). Fluorescently labeled secondary antibodies (Rockland IRDye 800) were diluted 1:5000 in a 1:1 solution of PBS and Odyssey blocking reagent containing 0.2% Tween and incubated at RT for 1h. The membranes were rinsed and visualized on an Odyssey Infrared Imaging System (Li-Cor Biosciences).
CAT enzyme levels in response to E2 treatment were analyzed by paired t-test compared to vehicle treatment. Quantitative RT-PCR for ERα mRNA in different cell types was analyzed by ANOVA followed by Fishers LSD post hoc test. p<0.05 was considered statistically significant. ERα and GFAP mRNA expression as well as cell counts obtained for each patient group were analyzed by one-way ANOVA followed by Fishers LSD post hoc analysis where appropriate. Correlation analysis between data obtained for HIV-infected patient groups and the severity of dementia was determined by Spearman nonparametric correlation. p<0.05 was considered statistically significant.
Transfection of Tat plasmid DNA into SVGA-CAT cells results in the constitutive expression of Tat protein. Since these cells are stably transfected with the HIV-LTR promoter driving the CAT reporter enzyme, subsequent binding of Tat to the TAR region of the HIV-LTR results in transcriptional activation and the production of CAT enzyme. CAT levels were measured by ELISA and were shown to increase in cells expressing Tat protein compared to cells transfected with an empty vector control lacking the Tat coding sequence (Fig. 1A). Similar to previously published data from our laboratory, we confirmed E2 (0.01–1nM) treatment of SVGA-CAT cells attenuated the Tat-induced activation of the LTR. In addition, E2 treatment of HIV-infected SVGA cells resulted in a decrease in the expression of the HIV matrix protein, p24, at 3 days up to 1 week postviral infection compared to vehicle control (Fig. 1B).
To determine whether E2 suppressed Tat-induced LTR activation through an ERα-dependent mechanism, endogenous expression of ERα in SVGA cells was investigated. ERα mRNA expression in SVGA cells compared to ERα-positive HEK293 cells and MCF-7 breast cancer cells was determined by qRT-PCR. Results showed ERα mRNA was detected in SVGA cells and HEK293 cells (Fig. 2A, left); however, in comparison to MCF-7 breast cancer cells, which express abundant levels of ERα mRNA and protein, the ERα mRNA levels detected in SVGA cells and HEK293 cells were negligible (Fig. 2A, middle). Furthermore, ERβ mRNA levels were detected in all three cell types but were even lower than ERα mRNA levels (Fig. 2A, right). Since ERα mRNA was detected in SVGA cells, albeit at low levels, endogenous expression of ERα protein in SVGA cells compared to ERα-negative HeLa and HEK293 cells was determined by immunocytochemistry (Fig. 2B) and Western blot (Fig. 2C). As shown, there was no clear evidence of ERα expression in SVGA, HeLa, or HEK293 cells compared to cells overexpressing ERα in which obvious nuclear localization of the protein was observed. Western blot analysis of ERα in nuclear and cytoplasmic fractions isolated from the various cell types as indicated (Fig. 2C) revealed a protein band ~66kDa (the molecular weight of ERα) in HEK293 cells overexpressing ERα protein. However, no ERα protein was detected in SVGA cells.
To investigate the role of ERα in E2 suppression of Tat-induced LTR activation, SVGA cells were transfected with plasmid DNAs encoding for Tat and ERα prior to CAT analysis. As expected, treatment of SVGA cells with 1nM E2 reduced Tat-induced CAT expression by ~40% compared to vehicle. However, surprisingly, the presence of ERα negated this effect and no differences in Tat-induced CAT expression were found following E2 treatment compared to vehicle. E2 treatment of SVGA cells in which ERα was overexpressed, in the absence of Tat, appeared to result in an increase in CAT expression compared to vehicle, but this was not significant (Fig. 3A). ERα protein levels were monitored by Western blot analysis (data not shown). To investigate whether the E2-mediated suppression of Tat-induced LTR activation occurs only in SVGA cells, HEK293 cells were transfected in the same way as SVGA cells with the exception that all HEK293 cells were also transfected with the LTR-CAT construct since the HEK293 cell line lacked a stably transfected LTR reporter. CAT assays showed that E2 had no effect on Tat-induced activation of the LTR compared to vehicle (Fig. 3B). In addition, the presence of ERα in the presence or absence of E2 had no effect on LTR activation.
To understand the significance of our findings for the human HIV-infected brain, we next measured the levels of ERα mRNA and protein in postmortem human hippocampal tissue. Quantitative real-time PCR of cDNA generated from human hippocampal tissue revealed no change in ERα mRNA levels between any of the patient groups (Fig. 4A). However, GFAP mRNA levels were shown to increase in tissue obtained from HIV-infected patients compared to uninfected controls (p<0.001). Interestingly, both MCMD and HAD groups showed a decline in GFAP mRNA compared to the HIV+ control group (p<0.05) with levels similar to those from uninfected patient samples (Fig. 4B). Analysis of the severity of cognitive decline following HIV infection was shown to negatively correlate with the level of GFAP mRNA in the hippocampus (p=0.005, Spearman's nonparametric correlation). This suggests a significant relationship between GFAP mRNA levels and neurological progression from HIV infection to HAD.
Human hippocampal sections were processed for immunocytochemical detection of ERα or GFAP protein. ERα was detected in pyramidal neurons of ammons horn as well as dentate granule cells and hilar neurons (data not shown). To determine whether hippocampal astrocytes express ERα, we performed localization analysis of immunofluorescent labeled sections. Images of GFAP (green), ERα (red), and DAPI (blue) labeling were captured. Images revealed the presence of numerous GFAP-IR cells positive for ERα in each patient group (Fig. 4C). A magnified image of a cell that was identified as an ERα-IR astrocyte is shown in Fig. 4D, in which cytoplasmic labeling of ERα is shown.
Cell counts of dentate granule cells that expressed both ERα and GFAP as well as total GFAP-IR cells revealed no difference in the density of ERα-GFAP-IR cells between groups (Fig. 5A), while the density of GFAP-IR cells increased in HIV+ cognitively normal patients compared to uninfected controls but was decreased in both MCMD and HAD compared to the HIV+ cognitively normal group (Fig. 5B). These findings are in agreement with the mRNA expression of ERα and GFAP we observed in the hippocampus. However, the density of ERα-GFAP-IR cells normalized to total GFAP-IR cells was decreased by 23% in the HIV+ control group compared to the HIV− group (Fig. 5C, p<0.05), but was increased in the HAD group compared to the HIV+ control group (p<0.05). In addition to the quantification analysis of the data, a significant correlation between the severity of dementia following HIV infection and the density of ERα-IR astrocytes was observed (p=0.0049, Spearman nonparametric correlation).
The principal finding of the present study was that E2 suppressed Tat-induced HIV LTR activation through an ERα-independent mechanism. Paradoxically, we found that the overexpression of ERα negated the E2 effect in reducing Tat transactivation in a manner that was cell type specific. In addition, the data showed an increased percentage of ERα-IR astrocytes in the hippocampus of individuals with HAD in comparison to cognitively normal HIV+ patients. Despite abundant data illustrating extensive astrogliosis in HIV brain and the increase in ERα expression within reactive astrocytes associated with neurodegenerative disease or injury, this is the first study to investigate changes in the expression and distribution of ERα in the hippocampus of HIV-infected patients and the relationship between ERα expression and cognitive function.
In general, estrogen action is mediated through the binding of either of the identified estrogen receptors, ERα and ERβ. The ERα agonist (PPT) mimics E2 in suppressing Tat transactivation, which suggests that ERα may be involved in Tat-mediated transcription.21 In the present study we found ERα mRNA in SVGA and HEK 293 cells, however, these levels were negligible compared to MCF-7 cells, which express abundant levels of ERα mRNA and protein. HEK293 cells express ERα mRNA but lack ERα protein38 whereas HeLa cells lack both ERα mRNA and protein. For this reason, HeLa and HEK293 cells were chosen as negative controls. Similarly, ERβ mRNA levels were analyzed by qRT-PCR and found to be negligible compared to the levels of ERα mRNA in SVGA cells. For this reason we did not further investigate a role for ERβ in mediating the effects of estrogen on attenuating HIV LTR activation in SVGA cells.
In agreement with reported observations no ERα protein was detected in HeLa or HEK293 cells by immunocytochemistry or Western blot. Interestingly, although SVGA cells appeared to have detectable levels of ERα mRNA, we failed to detect ERα protein in these cells. This result was surprising given that astrocytes and neurons express ERα and ERβ.35,40 This may be explained by the properties of SVGA cells compared to primary astrocytes. SVGA cells were originally derived from human fetal brain tissue that was transfected with a replication-deficient mutant of SV40.29 This resulted in the creation of an immortalized but not transformed cell line consisting of a homogeneous population of glial cells, identified by their immunoreactivity to GFAP.29 SVGA cells were derived from fetal human brain tissue isolated at 8–12 weeks gestation, a time when ERα expression in the cortex has started to decline, while expression in the hippocampus has yet to occur.37,41 Together this suggests that the absence of detectable levels of ERα in SVGA cells may be because the cells are not fully differentiated into mature astrocytes. Indeed, SVGA cells have been described as a presumptive progenitor cell of the CNS that morphologically does not appear differentiated.42 Furthermore, despite the astrocytic phenotype the cells adopt when cultured in 10% serum-containing media in the present studies, SVGA cells can alternatively be cultured as neuronal cells in the presence of N2-supplemented neurobasal media.42–44
Since many astrocytes in vivo express ERα, we investigated the effect of ERα overexpression in SVGA cells on the effect of E2 in decreasing Tat-mediated HIV promoter activation. Results showed that in the presence of ERα no E2-mediated decrease in LTR activation by Tat was observed. E2 treatment of cells transfected with ERα but not Tat DNA appeared to increase LTR activation but did not reach significance. Interestingly E2 had no effect in the absence or presence of ERα in HEK293 cells, indicating that the effect of E2 is cell type specific. The E2-mediated suppression of Tat-induced HIV LTR activation in astrocytes in the human brain would have tremendous therapeutic potential because this would lead to a decrease in HIV protein production and concomitant neuronal damage. To address the relevance of our data illustrating the effect of ERα in negating the effect of E2 we investigated the expression of ERα protein in astrocytes in the HIV-infected human brain and how this relates to cognitive function.
In response to injury mediated by trauma, ischemia, or neurodegenerative disease ERα expression in the brain is increased.22,35,45,46 ERα expression increases in reactive astrocytes in the rodent, primate, and human brain,24,27 and astrogliosis is a common pathological finding following injury to the CNS.47 With the use of GFAP immunolabeling, we identified reactive astrocytes in hippocampal samples obtained from each patient. GFAP-IR cells were detected mainly in the dentate gyrus, a pattern that has been reported for HIV-infected brains.48 As expected, GFAP expression was higher in the HIV+ samples compared to uninfected controls, indicative of astrogliosis, while in HIV+ samples derived from cognitively impaired patients, GFAP expression was decreased in the dentate gyrus compared to patients classified as cognitively normal. Interestingly, the decline in reactive astrocytes following the initial viral-induced increase in astrogliosis is correlated with neurological progression to HAD as shown by the directly proportional relationship between cognitive status in HIV and reduced GFAP expression. These findings are in agreement with a report describing a correlation between neurological progression and astrocyte apoptosis in HAD.49
Immunohistochemical analysis of the distribution of ERα and GFAP colabeled cells in the hippocampus showed that both the density of ERα-IR cells and ERα mRNA expression in the dentate gyrus did not differ between cognitively impaired and cognitively normal HIV-infected patients. This finding was unexpected and deserves further investigation since ERα expression is increased in the rat following ischemia50 and the human brain in Alzheimer's disease (AD).51 However, the percentage of ERα-GFAP-IR cells decreased in HIV-infected patients compared to uninfected controls and was increased in HIV+ patients with HAD compared to cognitively normal patients. The increase in the percentage of the ratio of ERα-GFAP-IR cells was progressive and was shown to correlate with the severity of cognitive impairment following HIV infection.
It has previously been reported that in the hippocampus of AD patients the density of ERα-IR astrocytes is increased while the percentage of ERα-IR astrocytes remains unaltered compared to nondemented control subjects.35 It is possible that the role of ERα expression in the hippocampus in neurological progression to HAD following HIV infection may differ from the development of dementia in AD. Of importance are the data supporting a role for estrogen neuroprotection in delaying the onset and progression of AD,52 which for HIV is not entirely clear at present. The increase in the ratio of ERα-IR astrocytes in the hippocampus of HAD patients may provide a desirable target for estrogen action. In support of this, an increase in ERα-IR astrocytes has been reported in the rat following a stab wound and excitotoxic injury24 and estrogen has been shown to provide neuroprotection in both of these injuries.53–55 Therefore, the fact that we did not observe the same trend in our HIV samples implies that ERα upregulation in astrocytes is not a general response to injury but appears to be injury specific.
Despite supporting evidence for E2 as a neuroprotective agent in the brain, gender differences in the development and severity of cognitive decline in HIV-infected women are not entirely clear; however, it has been reported that women on hormone replacement therapy were less likely to develop HAD than those who were not.56 Overall, too few studies have been conducted to determine a neuroprotective role for estrogen in HIV dementia; however, it is possible that E2 could play a protective role in the neuropathogenesis of HAD by attenuating the Tat-induced activation of the HIV promoter resulting in reduced HIV protein synthesis and concomitant neuronal damage. However, based on the data presented in this study, the time at which E2 would be most beneficial in the reduction of HIV transcription would be at early stages of disease progression when patients have yet to develop cognitive impairment (illustrated in Fig. 6). At this stage, HIV entry into the brain has occurred, the numbers of reactive astrocytes that do not express ERα far exceed those that do, and E2 treatment would significantly lower the levels of Tat-induced viral transcription. In contrast, a patient suffering HAD is unlikely to benefit from E2 treatment because even though astrocyte cell death leads to reduced levels of HIV proteins, the damage caused by the initial increase in reactive astrocytes has already occurred and a slight decrease in HIV protein expression at this stage is unlikely to provide adequate neuroprotection. Although we showed that E2 suppressed Tat-induced LTR activation in an astrocyte cell line, it remains to be determined whether this occurs in astrocytes in the human brain and furthermore whether endogenous ERα expression in astrocytes interferes with this effect.
Since our data suggest that ERα is not responsible for mediating the effects of E2 LTR activation, the mechanism behind the E2-mediated suppression of HIV transcription remains a mystery. The G-protein-coupled receptor, GPR30, is a recently identified receptor shown to bind E2 with high affinity leading to the activation of MAP kinase57and PI3 kinase pathways in MCF-7 cells.58 PI3K activation has been shown to inhibit Tat-induced LTR activation59 and GPR30 is expressed and is activated by E2 in SVGA cells (unpublished observations). Therefore a role for GPR30 in mediating the effects of estrogen in the attenuation of HIV transcription cannot be ruled out. Our data suggest that estrogen can act as a protective agent in the brain of HIV-infected individuals; however, more extensive clinical studies are needed to address gender differences in HIV dementia.
The authors would like to thank the National NeuroAids Tissue Consortium for supplying the postmortem human tissue. We would also like to thank Dr. S. Estus for supplying the GFAP primers used for qRT-PCR and help with RT-PCR data analysis, Dr. T. McClintock for help with RT-PCR experimental design and data analysis, Ms. C. Long for technical assistance with tissue sectioning, and Dr. R. Kryscio for help with statistical analysis. This work was supported by Grant P20 RR 15592 from the National Center for Research Resources (NCRR), a component of the National Institutes of Health (NIH).
No competing financial interests exist.