Christopher P. Anderson

Astrocytes and Microglia Respond to Estrogen with Increased apoE mRNAin Vivoandin Vitro

This study examined the regulation of apolipoprotein E (apoE) by 17beta-estradiol (E2) in brain glia, using rats with regular ovulatory cycles as an in vivo model and cultured astrocytes and mixed glia as in vitro models. Two brain regions were examined which had demonstrated transient synaptic remodeling during the estrous cycle. In the hippocampal CA1 region and the hypothalamic arcuate nucleus, apoE mRNA was elevated at proestrus when plasma E2 was high and synaptic density was increasing. Both astrocytes and microglia contributed to this increase in apoE mRNA. In vitro, E2 treatment had no effect on apoE mRNA levels in monotypic cultures of either astrocytes or microglia. In contrast, mixed glial cultures responded to E2 with increased apoE mRNA and protein, suggesting that heterotypic cellular interactions are important in the brain response to estrogens. In situ hybridization in combination with cell-specific markers showed that E2 increased apoE mRNA levels in both astrocytes and microglia. These results, which are the first evidence of apoE mRNA localization to microglia in vivo and the control of apoE expression in brain cells by estrogens, are discussed in terms of the possible protective role of E2 in Alzheimers disease and prior findings that emphasize the expression of apoE mRNA in astrocytes within the brain.

Glial Fibrillary Acidic Protein Transcription Responses to Transforming Growth Factor‐β1 and Interleukin‐1β Are Mediated by a Nuclear Factor‐1‐Like Site in the Near‐Upstream Promoter

Abstract: Elevated expression of glial fibrillary acidic protein (GFAP) is associated with astrocyte activation during responses to injury in the CNS. Because transforming growth factor‐β1 (TGF‐β1) and interleukin‐1β (IL‐1β) are released during neural responses to injury and because these cytokines also modulate GFAP mRNA levels, it is of interest to define their role in GFAP transcription. The increases of GFAP mRNA in response to TGF‐β1 and decreases in response to IL‐1β were shown to be transcriptionally mediated in rat astrocytes transfected with a luciferase‐reporter construct containing 1.9 kb of 5′‐upstream rat genomic DNA. Constructs containing sequential deletions of the rat GFAP 5′‐upstream promoter identified a short region proximal to the transcription start (‐106 to ‐53 bp) that provides full responses to TGF‐β1 and IL‐1β. This region contains an unusual sequence motif with overlapping nuclear factor‐1 (NF‐1)‐ and nuclear factor‐κB (NF‐κB)‐like binding sites and homology to known TGF‐β response elements. Mutagenesis (3‐bp exchanges) in ‐70 to ‐68 bp blocked the induction of GFAP by TGF‐β1 and the repression by IL‐1β. Gel shift experiments showed that the DNA segment ‐85 to ‐63 bp was bound by a factor(s) in nuclear extracts from astrocytes. The concentrations of these DNA binding factors were increased by treatment of astrocytes with TGF‐β1 and decreased by IL‐1β. Binding of these nuclear factors was blocked by mutation of ‐70 to ‐68 bp. Despite homology to NF‐1 or NF‐κB binding sites in the GFAP promoter at segment ‐79 to ‐67 bp, anti‐NF‐κB or anti‐NF‐1 antibodies did not further retard the gel shift of the nuclear factors/DNA complex. Moreover, astrocytic nuclear proteins do not compete for the specific binding to NF‐1 consensus sequence. Thus, nuclear factors from astrocytes that bind to the ‐85‐ to ‐63‐bp promoter segment might be only distantly related to NF‐1 or NF‐κB. These findings are pertinent to the use of GFAP promoter constructs in transgenic animals, because cis‐acting elements in the GFAP promoter are sensitive to cytokines that may be elaborated in response to expression of transgene products.

Effects of Age on Gene Expression during Estrogen-Induced Synaptic Sprouting in the Female Rat ☆

Age and estrogen treatment influenced fiber outgrowth and compensatory neuronal sprouting after unilateral entorhinal cortex lesions (ECL) which model Alzheimer disease-like deafferentation in the dentate gyrus of the hippocampus. In young F344 rats (3 months old), ovariectomy (OVX) decreased reactive fiber outgrowth by 60%. Sprouting in middle-aged rats (18 months old) was reduced in intact females; no further reduction was caused by OVX. Several astrocyte mRNAs were measured in the dentate gyrus of young and middle-aged female rats in three different estrogen states (sham OVX, OVX, or OVX + estradiol) 1 week after ECL. Glial fibrillary acidic protein (GFAP) mRNA was twofold greater in middle-aged rats than young, although both ages showed threefold increases in response to ECL. In prior studies GFAP was found to be decreased by estradiol treatment 3-4 days after ECL; in this study GFAP mRNA had returned to sham OVX levels in young rats by 7 days post-ECL. Surprisingly, estradiol treatment increased GFAP mRNA levels by 25% above OVX in middle-aged rats. Apolipoprotein E (apoE) mRNA was decreased 20% by age in the dentate, although both age groups showed a 25% increase in apoE mRNA in response to ECL. Apolipoprotein J (apoJ) mRNA was increased 20% in the dentate gyrus of middle-aged rats, and both age groups responded to ECL with a 65% increase in apoJ mRNA. The estrogen state did not alter levels of either apolipoprotein mRNA in the deafferented dentate. The data suggest that the estrogen-induced decrease of GFAP in response to lesions does not persist at 7 days post-ECL during sprouting. Overall effects of age on the dentate gyrus include elevated GFAP mRNA and decreased apoE mRNA. The cortical wound site showed consistent enhancement of GFAP mRNA in both age groups by estradiol above sham OVX and greater responses in middle-aged rats.

Variations of synaptotagmin I, synaptotagmin IV, and synaptophysin mRNA levels in rat hippocampus during the estrous cycle.

Periodic changes in ovarian steroid levels during fertility cycles affect learning both in humans and in rats in parallel with electrophysiological and morphological fluctuations in selective neuronal populations. In particular, during the estrous cycle of the female rat, hippocampal CA1 region undergoes cyclic modifications in synaptic density. To investigate the molecular mechanisms involved in synaptic remodeling during the estrous cycle, we analyzed the expression of three presynaptic markers, synaptotagmin I, synaptotagmin IV, and synaptophysin, in the female adult rat brain by in situ hybridization. Relative abundance in mRNA for these three markers was quantified at four phases of the estrous cycle: diestrus, proestrus (AM and PM), and estrus. mRNA levels for syt1 exhibited cyclic variations in pyramidal neurons of the CA3 region of hippocampus during the estrous cycle, while mRNA levels for syt4 and SYN were relatively invariant in this or other regions of the hippocampus. Because CA3 pyramidal neurons make synaptic contacts in CA1, modulation of syt1 expression in CA3 may participate in the changes in synaptic density observed in CA1 during the estrous cycle. Furthermore, both syt1 and SYN mRNA varied cyclically in layer II, but not in layer III of entorhinal cortex, while syt4 remained unchanged throughout the cycle. These data suggest that regular variations in steroid hormone levels during fertility cycles may alter the properties of several networks involved in information processing and learning and memory through altered levels of presynaptic proteins.

Methylation of the glial fibrillary acidic protein gene shows novel biphasic changes during brain development

The gene for glial fibrillary acidic protein (GFAP) was analyzed in the rat for developmental changes in methylation of cytosine at CpG sequences as a correlate of the onset of GFAP mRNA expression and for the effect of methylation on GFAP promoter activity. The methylation of nine CpG sites in the GFAP promoter and ten sites in exon 1 was analyzed in F344 rats by a quantitative application of ligation‐mediated polymerase chain reaction. Whole rat brain poly(A) + RNA showed an exponential increase of GFAP mRNA after embryo day 14 that reached stable adult levels by postnatal day 10. During development, only the seven CpG sites in the far‐upstream promoter showed large changes in methylation; these sites constitute the brain‐specific domain of methylation described in adult rats (Teter et al: J Neurosci Res 39:680, 1994). These seven CpG sites showed a cycle of demethylation during the onset of GFAP transcription in the embryo (between embryonic day 14 and postnatal day 10) followed by remethylation at later postnatal ages when GFAP mRNA remains prevalent. The minimum levels of methylation across these CpG sites displayed a gradient with the lowest minima at the 3′ sites. This demethylation/remethylation cycle is a novel phenomenon in DNA methylation during perinatal development. The demethylation/remethylation cycle during development was also shown by the opposite‐strand cytosines. Two cytosines in this region that are conserved in rat and mouse also undergo the same demethylation/remethylation cycle in the mouse GFAP gene during development, implying evolutionary conservation and functional significance. As a further test of functional significance, a Luciferase reporter gene assay was evaluated in primary cultured astrocytes; the activity of the GFAP promoter was reduced when it was methylated at one or all CpG sites. Therefore, the GFAP promoter may be activated in rodent development by transient demethylation of a conserved brain‐specific methylation domain.

Equine estrogens induce apolipoprotein E and glial fibrillary acidic protein in mixed glial cultures.

Premarin, which contains several equine estrogens, as well as estradiol (E2) as a minor component, is widely used for replacement therapy of estrogen deficits, but little is known of its direct actions on brain cells. In mixed glial cultures, apolipoprotein E (apoE) and glial fibrillary acidic protein (GFAP) are induced by estrogens. GFAP induction showed an inverted-U shape E2 dose response, with a maximum induction at 1 pM, whereas apoE mRNA induction was greatest at 100 pM. GFAP and ApoE mRNAs were induced by equine estrogens in the following order: E2=equilin>estrone>17 alpha-dihydroequilenin. However, the induction of apoE secretion by 17 alpha-dihydroequilenin was as effective as by the other estrogens. The greater response of apoE secretion than GFAP mRNA induction to 17 alpha-dihydroequilenin might be therapeutically important because of the glial scarring during brain lesions, in which GFAP induction has a major role in inhibiting neurite outgrowth, whereas apoE secretion supports neurite outgrowth.

Failure of dietary restriction to retard age-related neurochemical changes in mice

Age-sensitive neurochemical measures and estrous cyclicity were studied in female mice from the long-lived C3B10F1 strain fed either a control diet or subjected to dietary restriction (DR) from 3 weeks of age. Striatal dopaminergic D2 receptor density decreased by 25% from 9-10 months to 28-30 months of age in the control group. This decline was uninfluenced by DR. Anterior pituitary dopamine + dihydroxyphenylacetic acid content increased by 2.5 fold with age in the control group but DR failed to oppose this age-related change. In contrast to DRs lack of influence on these two neurochemical measures were findings on estrous cyclicity. Although mice on DR did not display estrous cycles, cyclicity was rapidly initiated when these mice were switched to the control diet at 12 and even at 22 months of age. Thus, limited aspects of neuroendocrine aging were retarded by DR in this long-lived mouse model.

Aging and increased hypothalamic glial fibrillary acid protein (GFAP) mRNA in F344 female rats. Dissociation of GFAP inducibility from the luteinizing hormone surge.

During reproductive aging, female rodents show impaired inducibility of the estradiol (E2)-induced (preovulatory) surge of luteinizing hormone (LH), which is associated with hypothalamic neuronal impairments of aging. To evaluate if astrocytes show comparable age changes, we analyzed the regulation of glial fibrillary acidic protein (GFAP) mRNA which is transiently increased in the arcuate nucleus in association with the proestrus LH surge in young rats. In aging (18-month-old) F344 rats in persistent estrus, the loss of the E2-induced LH surge was paralleled by the lack of increased GFAP mRNA in the arcuate nucleus (in situ hybridization with X-ray film). We then tested the hypothesis that restoration of the LH surge by chronic ovariectomy (OVX) in aging rats would restore or block GFAP mRNA induction, respectively. Despite restoration of the inducible LH surge in aging rats by chronic OVX, there was no induction of GFAP mRNA in the arcuate nucleus. Moreover, young rats given chronic E2 implants for 6 weeks as a model for persistent estrus, retained induction of arcuate nucleus GFAP mRNA, despite loss of the induced LH surge. The aging rats were highly sensitive to gross pituitary enlargement from chronic E2 with 50% mortality; thus, the F344 genotype is not optimum for studies of aging-E2 interactions that require prolonged E2 treatments. More detailed cellular level analysis of GFAP mRNA is needed to define the relationship of GFAP expression to synaptic reorganization during the LH surge. The aging rats also showed higher levels of GFAP mRNA in the arcuate nucleus and ventromedial nucleus, consistent with the general trend for elevated GFAP mRNA during aging in other brain regions and in both sexes.

Polygenic influences on the length of oestrous cycles in inbred mice involve MHC alleles

Genetic influences on female reproductive cycles were analysed in histocompatibilitycongenic strains of mice. Oestrous cycles of young, virgin mice of inbred‐congenic strains, hybrid crosses (F1, and parental‐hybrid backcrosses (F2) were monitored for 3 months. Oestrous cycles were categorized by length (inter‐oestrous interval): 4,5,6, or 7–14 days. Mice with the following H‐2 haplotypes had a greater proportion of 5‐day oestrous cycles: H‐2b, H‐2r, H‐2h2, H‐2h4, and H‐2i5. In contrast, the H‐2k and H‐2d haplotypes had mostly 4‐day oestrous cycles. Influences of H‐2 haplotype were seen on two genetic backgrounds, C57BL/10Sn and C3H. Non‐H‐2 alleles were also implied by different patterns of cycles between strains with the same H‐2b haplotype: C57BL/10Sn with predominantly 5‐day cycles vs. C57BL/6J with a mix of 4‐ and 5‐day cycles.

Hyperprolactinemia fails to increase striatal dopamine receptors in male and female C57BL/6J mice

Mice were administered prolactin by either homotypic anterior pituitary grafts (4 weeks) or subcutaneous osmotic pumps containing ovine prolactin (1 week). The pituitary grafts produced considerably greater elevations of circulating prolactin than the osmotic pumps. In neither case was any change detected in striatal [3H]spiperone binding to dopamine receptors. These results are discussed in the context of contradictory evidence obtained with rats.