H. Zheng and H. Xu contributed equally to this paper.
Modulation of Aβ peptides by estrogen in mouse models
Article first published online: 28 DEC 2001
Journal of Neurochemistry
Volume 80, Issue 1, pages 191–196, January 2002
How to Cite
Zheng, H., Xu, H., Uljon, S. N., Gross, R., Hardy, K., Gaynor, J., Lafrancois, J., Simpkins, J., Refolo, L. M., Petanceska, S., Wang, R. and Duff, K. (2002), Modulation of Aβ peptides by estrogen in mouse models. Journal of Neurochemistry, 80: 191–196. doi: 10.1046/j.0022-3042.2001.00690.x
- Issue published online: 28 DEC 2001
- Article first published online: 28 DEC 2001
- Received 3 July, 2001; revised manuscript received 16 October, 2001; accepted 17 October, 2001.
- Alzheimer's disease;
- transgenic mouse
Clinical studies have shown that estrogen deprivation through menopause is a risk factor in both the initiation and progression of Alzheimer's disease (AD) and that estrogen replacement therapy may be protective. One of the major pathological features in the human AD brain is the senile plaque, a proteinaceous structure composed mainly of heterogeneous peptides collectively known as A-beta (Aβ). In vitro studies have linked estrogen with Aβ modulation, suggesting that one-way that estrogen depletion at menopause may exacerbate the features of AD is through Aβ accumulation. To test this, two studies were performed on transgenic models of amyloidosis. Firstly, transgenic mice without detectable amyloid aggregates were subjected to ovariectomy and estradiol supplementation, and Aβ levels were assessed. Secondly, the effects of estrogen modulation were assessed in mice at an age when plaques would be forming initially. Overall, Aβ levels were higher in estrogen-deprived mice than intact mice, and this effect could be reversed through the administration of estradiol. These data suggest that, in vivo, estrogen depletion leads to the accumulation of Aβ in the CNS, which can be reversed through replacement of estradiol. These results provide evidence that post-menopausal estrogen depletion may be linked to an increased risk of AD through Aβ modulation.
APP, amyloid precursor protein
relative peak intensity
Estrogen deprivation has been implicated in the pathogenesis of Alzheimer's disease (AD) as there is strong evidence that post-menopausal estrogen replacement therapy (ERT) correlates with a reduced risk and delay in the onset of the disease (Tang 1996). Biochemical studies have shown that the estrogen analog 17β-estradiol (E2), affects the processing of the amyloid precursor protein (APP) in cultured cells (Jaffe et al. 1994; Chang et al. 1997; Xu et al. 1998) and can reduce the amount of the amyloidogenic fragments (i.e. Aβ40 and Aβ42) that are generated (Xu et al. 1998). Aβ is the main component of one of the cytopathogenic features in the human AD brain, the amyloid plaque, and accumulation of these peptides has been strongly implicated in AD through genetic and biochemical analyses. Thus, it is plausible that estrogen deprivation following menopause may increase the risk of developing AD through the accumulation of Aβ. Transgenic mice that generate high levels of human Aβ40 and Aβ42 develop amyloid deposits in the brain and, concomitantly, display several other features of human AD (Games et al. 1995; Hsiao et al. 1996). Several lines of transgenic mice generate Aβ at a high enough level to deposit amyloid into plaques that strongly resemble those seen in the human AD brain. These include mice that only over-express mutant APP such as the PDAPP line (Games et al. 1995) and Tg2576 (Hsiao et al. 1996), and lines of mice expressing both mutant APP and presenilin (PS1) transgenes (Borchelt et al. 1997; Holcomb et al. 1998; Lamb et al. 1999; McGowan et al. 1999). Here we show that ovariectomy of transgenic mice leads to accumulation of both soluble and insoluble Aβ peptides in vivo, suggesting a direct link between estrogen-depletion and exacerbation of an Alzheimer's-associated phenotype. In addition, the accumulation of Aβ peptides in response to ovariectomy can be reversed by estradiol treatment, suggesting that ERT may have therapeutic value in reducing the risk (or severity) of AD.
Materials and methods
Mouse surgery and analysis
Line Tg2576 over-expresses a mutant APP transgene and has elevated levels of human Aβ40 and Aβ42 relative to mouse, but does not deposit amyloid until 10–12 months of age (Hsiao et al. 1996). Double transgenic progeny from a cross between Tg2576 and a mutant PS1 line (Line PS1 5.1, M146L, Duff et al. 1996) have elevated levels of Aβ42 relative to the Tg2576 parent and form visible amyloid aggregates at approximately 16 weeks of age (Holcomb et al. 1998).
Tg2576 mice (termed ‘APP’) were used to study the effects of ovariectomy and estradiol supplementation on soluble Aβ40 and Aβ42 levels by enzyme-linked immunosorbant assay (ELISA) and mass spectrometry. Ovariectomy was performed at 13 weeks of age and sham-surgery was performed on littermate animals to provide a control group. For estrogen replacement groups, ovariectomized animals were simultaneously implanted subcutaneously with 90-day-release form of 17β-estradiol pellet at either 1.7 mg or 5 mg per pellet (Innovative Research of America, Sarasota, FL, USA). Mice were killed at 7 months of age. Prior to perfusion, uteri were weighed and plasma was collected for estradiol measurements using a radioimmunoassay (RIA) kit provided by Diagnostic Systems Laboratories, Inc. (Webster, TX, USA) according to the manufacturer's suggested method.
Double transgenic mice (termed ‘PS/APP’) were used to study the effect of ovariectomy on amyloid accumulation in the brain by ELISA. Ovariectomy was performed at 5 weeks of age and sham-surgery was performed on littermate animals. The animals were killed at 18 weeks of age, brains were removed for analysis and uterine weights were assessed. For estrogen supplementation, mice were ovariectomized at 5 weeks of age. At 11 weeks of age, estradiol was administered in the drinking water according to the method of Gordon et al. (1986). Briefly, estradiol (Sigma, St Louis, MO, USA) was dissolved in ethanol to a concentration of 5 mg/mL, then aliquoted into fresh drinking water at 5 µg/mL. The amount of water used was monitored twice weekly and fresh estradiol solution was administered. Four ovariectomized PS/APP mice received vehicle (water plus 0.1% ethanol), whereas three received estradiol in 0.1% ethanol. In all experiments, mice were maintained on a casein-based diet following ovariectomy to prevent the introduction of extraneous estrogen through soy-based mouse chow.
Formic acid extraction was used to solubilize all Aβ peptides in mouse hemibrains according to a method described in Refolo et al. (2000). A sandwich ELISA kit from BioSource International (Hopkinton, MA, USA) was used to measure Aβ40 and Aβ42 peptides according to the manufacturer's protocol in both PS/APP and APP mouse brain homogenates.
Immunoprecipitation/mass spectrometry (IP/MS) analysis of Aβ peptides was performed essentially as described in Wang et al. (1996). Aβ peptides were immunoprecipitated from formic acid extracted lysate with monoclonal antibody 4G8 (Senetek, Napa, CA, USA) using Protein G-Plus/Protein A-agarose beads (Oncogene Science, Inc., Cambridge, MA, USA) and analyzed by a matrix-assisted laser desorption/ionization time-of-flight mass spectrometer (MALDI-TOF-MS) (Voyager-DE STR BioSpectrometry Workstation, PerSeptive Biosystem). Synthetic Aβ 12–28 peptide (20 nm, Sigma) was added as internal standard. Aβ peptides terminating at amino acids 40 and 42 were identified by molecular mass and quantified using relative peak intensities. Each mass spectrum was averaged from 256 measurements and calibrated using bovine insulin as an internal mass calibrant. Relative peak intensities of Aβ peptides to internal standard were used for quantitative analysis.
Assessment of human APP and PS1 levels
Human APP (APPfl) and the C-terminal fragment of PS1 were detected by quantitative western blotting. Aliquots of formic acid extracts were lyophilized, resuspended in 60% acetonitrile and lyophilized again. The lyophilized proteins were then resuspended in 2% sodium dodecyl sulfate/phosphate-buffered saline (SDS/PBS), protein concentration was determined and duplicate 50 µg aliquots from each brain were resolved on 8% (APP) or 12% Tris–glycine gels (PS1 fragments). APPfl in the Tg2576 line was detected using the monoclonal antibody 6E10 (Senetek). Antibodies C1.61 (P.␣Mathews, Nathan Kline Institute) and α-loop PS1 (G. Thinakaran, University of Chicago) were used for the assessment of APP and the C-terminal fragment of PS1, respectively, in the PSAPP mice. sAPPα was detected by immunoprecipitation/western blotting as follows. Duplicate 200 µg aliquots from each brain were diluted five fold in IP buffer (1% Tx100/PBS). APPfl was depleted from the IP mixture by one round of immunoprecipitaion with 369 (S. Gandy, Nathan Kline Institute); the IP mixtures were then subjected to immunoprecipitation with antibody 6E10 (Senetek). The immunoprecipitated material was resolved on 6% Tris–glycine gels and immunoblotted with antibody 6E10. Blots were prepared in duplicate and multiple exposures from both blots were quantified by densitometry using the ScanAnalysis software.
Analyses of Aβ were performed using a non-parametric Wilcoxon analysis (Mann–Whitney U-test) using an SAS system NPAR1WAY procedure. Two sided p-values result from an ‘exact test.
To examine the effects of estrogen depletion and estrogen supplementation on soluble and insoluble Aβ peptide levels in the brains of transgenic mice, female mice from the APP and PS/APP lines were ovariectomized and a subgroup of the APP mice were then implanted with different concentrations (1.7 or 5 mg/pellet) of 17β-estradiol, whereas the PS/APP mice were supplemented with estradiol in the drinking water. The 1.7 mg/pellet provided close to physiological levels of circulating estrogen, whereas the 5 mg/pellet was super-physiological. In all cases, the ovariectomized mice showed complete atrophy of the uterus (mean weight < 0.01 g), indicating that the mice were gonadal steroid-depleted. Uterine weights in the sham-treated group and mice supplemented with estradiol were within the normal range (mean = 0.2 g). Circulating estradiol in the APP line was reduced from a mean value of 7.4 pg/mL (SD ± 5) in sham-treated animals to 4.3 (SD ± 1.6) in ovariectomized animals. In APP mice treated with a low dose of estradiol, the mean level of circulating estradiol rose to 24 pg/mL (SD ± 19) and in the high dose group, the mean level of circulating estradiol was 572 pg/mL (SD ± 35). Although the levels of estradiol in the CSF were not assessed, serum estradiol levels correlated with CNS Aβ levels. As expected, mice showed variance in the serum levels of estradiol as they were at different points in the estrus cycle at time of sacrifice.
Accumulation of Aβ peptides in APP mice
Analyses were performed with the investigator blind to the mouse treatment group in all cases. Ovariectomy of APP mice resulted in a significant increase in the level of soluble Aβ40 compared to sham-operated littermates (p < 0.05) and a trend towards increase for the Aβ42 peptides (Table 1). This increase was not associated with deposition into immunodetectable amyloid plaques as immunostaining using an Aβ-specific antibody (4G8) was negative (data not shown). In addition, ovariectomy of APP mice did not alter the level of APPfl (data not shown).
|Effect of ovariectomy and estrogen replacement on Aβ levels in APP mice||Number||Aβ40 (pmol/g protein, mean ± SEM)||Aβ42 (pmol/g protein, mean ± SEM)|
|APP Sham||7||70 ± 20||11 ± 2|
|APP OVX||10||85 ± 10a||14 ± 10|
|APP OVX + 1.7 mg E2||6||70 ± 10||3 ± 1b|
|APP OVX + 5 mg E2||7||50 ± 8b||Below limit of detection|
|Effect of ovariectomy and estrogen replacement on Aβ levels in PS/APP mice|
|PS/APP Sham||5||224 ± 42||149 ± 42|
|PS/APP OVX||8||337 ± 41||345 ± 53a|
|PS/APP OVX + vehicle||3||350 ± 65||491 ± 14|
|PS/APP OVX + estradiol||3||278 ± 33||237 ± 53a|
When the APP mice were administered a low dose of E2 following ovariectomy, the level of soluble Aβ40 was reduced to the level of sham-operated littermates. The reduction in the level of Aβ42 was more robust, however (p < 0.01). High doses of E2 following ovariectomy led to a highly significant reduction in the level of both Aβ40 and Aβ42 (p < 0.01 for Aβ40, Aβ42 fell below the detectable level). Data from IP/MS essentially confirmed the trends in the ELISA assays and a representative spectrogram is shown in Fig. 1.
Accumulation of Aβ peptides in PS/APP mice
By 18 weeks of age, the PS/APP mouse line used has been shown to develop immunodetectable amyloid deposits and this correlates well with ELISA measurements of Aβ peptides extracted in formic acid (Holcomb et al. 1998). Estrogen-depleted animals showed a significant increase in the mean level of peptides terminating at Aβ42 by ELISA (p < 0.05). Although there was a trend towards an increase for Aβ40 also, the increase was not statistically significant (p = 0.06) (Table 1 and Fig. 2).
To establish whether amyloid accumulation was related to transgene up-regulation, we examined brain lysates from PS/APP mice for full-length human APP (APPfl) and secreted APP (sAPPα), and for PS1 C-terminal fragment (PS1 CTF). Overall, there was no significant difference between the levels of APPfl, sAPPα or PS1 CTF in ovariectomized mice compared to sham-operated controls (Table 2).
|Sham||5||51.2 (1.56)||1.45 (0.08)||104.8 (0.7)|
|OVX||7||52.6 (1.88)||1.36 (0.12)||101 (1.98)|
Estradiol administered in drinking water to ovariectomized PS/APP mice resulted in a decrease in Aβ42 levels which reached borderline statistical signficance using our highly stringent two-tailed non-parametric test␣(p = 0.05).␣Reduction in Aβ40 levels was not significant (p = 0.4) (Table 1). This route of administration is considered to mimic human estradiol dosing as mice undergo diurnal variation in estradiol dose with the highest dose being administered when the mice are most active.
Data from both APP (Tg2576) and PS/APP (PS1.5.1/Tg2576) mice show that estradiol deprivation due to ovariectomy significantly affects the level of Aβ peptides in the brains of transgenic mice. Tg2576 mice at 7 months of age do not have appreciable amounts of insoluble, aggregated Aβ in the brain, and in this respect they are similar to the guinea pig model used by Petanceska et al. (2000). Both the guinea pig model and Tg2576 mice showed that soluble Aβ accumulates in estrogen-deprived animals, and this accumulation can be reversed by the administration of E2. As Tg2576 mice age, they begin to accumulate aggregated Aβ. Recently, Callahan et al. (2001) have shown that aged female Tg2576 mice have a significantly higher amyloid load than male littermates, suggesting that estrogen depletion as a result of the equivalent of mouse menopause impacts amyloid deposition in APP mice. We have now shown that mice with an accelerated amyloid phenotype (PS/APP) have significantly elevated Aβ in response to ovariectomy-induced estrogen depletion, and this elevation can be reversed following administration of E2. Data from both the guinea pig study and the PS/APP mice show an increased response of Aβ42 relative to Aβ40, suggesting that the mechanisms involved in the generation or clearance of Aβ42 are the preferential target for estradiol action. Data from the APP mice show that Aβ40 is significantly affected by ovariectomy, but the levels of Aβ42 were too variable to show a statistically significant increase. A trend was discernible, however. Data for E2 supplementation clearly showed that in common with the guinea pig model, E2 administration reduced the levels of Aβ42 in both the APP and PS/APP mouse models. Reduction of Aβ40 was less dramatic, and did not reach statistical significance in the PS/APP mouse. Mechanisms by which Aβ accumulates in the estrogen-depleted PS/APP mice likely include effects on APP processing, but also possibly effects on microglial clearance of amyloid once deposition is initiated, as estrogen has been shown to induce glia to take up aggregated Aβin vitro (Li et al. 2000).
Analysis of protein levels in the mice suggests that estrogen deprivation does not lead to up-regulation of the promoter of either the APP transgene or the PS1 transgene as the levels of APPfl and PS1 CTF and are statistically unchanged, and in fact show a trend towards decline in ovariectomized mice compared to sham-operated animals (Table 2). Thus, the increase in Aβ seen in the ovariectomized animals cannot be attributed to increased levels of APP protein, or mutant PS1 CTF. Data from cultured cells exposed to E2 suggest that estrogen affects APP processing such that soluble Aβ levels decrease and, concomitantly, sAPPα levels increase (Xu et al. 1998). Estrogen-deprived mice might therefore be expected to show an increase in Aβ and a decrease in sAPPα. Although there was a trend towards decline of the sAPPα fragment in the ovariectomized PS/APP mice, no significant difference was seen for this fragment. Data from APP mice were inconclusive. Lack of correlation between cell culture and in vivo experiments in terms of the effect of Aβ modulation on sAPPα has also been noted in other studies (Savage et al. 1998).
Overall, therefore, it appears that estrogen modulates the accumulation of Aβ in the CNS of transgenic mice. As estrogen has been associated with synaptogenesis (McEwen and Woolley 1994) and cognitive function (Luine 1985), it will be interesting to see if long-term deprivation of gonadal hormones exacerbates two other Aβ-related features of these mouse models: disorganization of cortical cholinergic neurons (Wong et al. 1999) and cognitive impairment (Holcomb et al. 1998). In general, our observations may explain why women undergoing estrogen replacement therapy have a reduced risk for developing AD, and provides a model system in which the interplay of Aβ and estrogen, and the efficacy of estrogen-based therapeutics, can be tested in vivo.
The authors wish to thank Cindy Yu (Mayo) and Brian Malester (NKI) for technical assistance, Eugene Laska and Joe Walderling (NKI, Department of Biostatistics) for statistical analyses and Karen Hsiao-Ashe (University of Minnisota) for the gift of Tg2576 mice. This work was supported by NIH grants AG10485 to KD/JS, AG09464 to SP and Altzheimer’s Association grants to HZ, HX, SP, RW and LR.
- 1997) Accelerated amyloid deposition in the brains of transgenic mice coexpressing mutant presenilin 1 and amyloid precursor proteins. Neuron 19, 939–945. , , , , , , , and (
- 2001) Augmented senile plaque load in aged female β-amyloid precursor protein transgenic mice. Am. J. Pathol. 158, 797–801. , , , , and (
- 1997) Estrogens influence growth, maturation, and amyloid beta-peptide production in neuroblastoma cells and in a β-APP transfected kidney 293 cell line. Adv. Exp. Med. Biol. 429, 261–271. , and (
- 1996) Increased amyloid-β42 (43) in brains of mice expressing mutant presenilin 1. Nature 383, 710–713. , , , , , , , , , , , , , , , and (
- 1995) Alzheimer-type neuropathology in transgenic mice overexpressing V717F β-amyloid precursor protein. Nature 373, 523–527. , , , , , , , , , et al. (
- 1986) Effective oral administration of 17β-estradiol to female C57BL/6J mice through the drinking water. Biol. Reprod. 35, 1088–1095. , , and (
- 1998) Accelerated Alzheimer-type phenotype in transgenic mice carrying both mutant amyloid precursor protein and presenilin 1 transgenes. Nat. Med. 4, 97. , , , , , , , , , , , , , , , , and (
- 1996) Correlative memory deficits, Aβ elevation, and amyloid plaques in transgenic mice. Science 274, 99–102. , , , , , , and (
- 1994) Estrogen regulates metabolism of Alzheimer amyloid β precursor protein. J. Biol. Chem. 269, 13065–13068. , , and (
- 1999) Amyloid production and deposition in mutant amyloid precursor protein and presenilin-1 yeast artificial chromosome transgenic mice. Nat. Neurosci. 2, 695–697. , , , , , , and (
- 2000) Estrogen enhances uptake of amyloid β-protein by microglia derived from the human cortex. J. Neurochem. 75, 1447–1454. , , , , and (
- 1985) Estradiol increases choline acetyltransferase activity in specific basal forebrain nuclei and projection areas of female rats. Exp. Neurol. 89, 484–490. (
- 1994) Estradiol and progesterone regulate neuronal structure and synaptic connectivity in adult as well as developing brain. Exp. Gerontol. 29, 431–436. and (
- 1999) Amyloid phenotype characterization of transgenic mice over-expressing both mutant amyloid precursor protein and mutant presenilin 1 transgenes. Neurobiol. Dis. 6, 231–244.DOI: 10.1006/nbdi.1999.0243 , , , , , , , , , and (
- 2000) Ovariectomy and 17β-estradiol modulate the levels of Alzheimer's amyloid β peptides in brain. Neurology 34, 2212–2217. , , and (
- 2000) Hypercholesterolemia accelerates the Alzheimer’s amyloid pathology in a transgenic mouse model. Neurobiol. Dis. 7, 321–331. , , , , , , , and (
- 1998) Turnover of amyloid beta-protein in mouse brain and acute reduction of its level by phorbol ester. J. Neurosci. 18, 1743–1752. , , , , , , , and (
- 1996) Effect of oestrogen during menopause on risk and age at onset of Alzheimer's disease. Lancet 348, 429–432. (
- 1996) The profile of soluble amyloid β protein in cultured cell media. Detection and quantification of amyloid beta protein and variants by immunoprecipitation-mass spectrometry. J. Biol. Chem 271, 31894–31902. , , and (
- 1999) Reorganization of cholinergic terminals in the cerebral cortex and hippocampus in transgenic mice carrying mutated presenilin-1, and amyloid precursor protein transgenes. J. Neurosci. 19, 2706–2716. , , and (
- 1998) Estrogen reduces neuronal generation of Alzheimer β-amyloid peptides. Nat. Med. 4, 447–451. , , , , , , , , , , , , , , , , , and (