Lack of an Association of Estrogen Receptor α Gene Polymorphisms and Transcriptional Activity With Alzheimer Disease

Background  Long-term cognitive decline in postmenopausal women is associated with aging and Alzheimer disease (AD). Estrogen replacement therapy has been reported to reduce the risk of developing AD. The distribution of estrogen receptors (ERs) in neurons overlaps that of the brain neurons known to develop AD. Estrogen increases the secretion and metabolism of amyloid precursor protein, may help synapse formation, and is reported to protect neurons from toxins. Restriction fragment length polymorphisms (RFLPs) of the ERα gene at intron 1 and exon 2 were associated with a low bone mineral density in postmenopausal women and also with AD in a Japanese population.

Objective  To determine whether ERα gene polymorphisms are associated with transcriptional activity and AD.

Methods  A luciferase reporter assay analyzed enhancer activity of the ERα gene at intron 1 and exon 2. This activity was evaluated according to the RFLPs. The RFLPs of the ERα gene were determined in Japanese patients clinically diagnosed as having AD, white patients diagnosed as having AD at autopsy, and corresponding healthy control subjects. The RFLPs were also evaluated for the contribution of the ERα gene RFLPs to AD.

Results  We found weak (about 2-fold) enhancer activity of the ERα gene, which differed among RFLPs. Although there were racial differences in these polymorphisms, we could not confirm the previously reported association between ERα gene polymorphisms and AD.

Conclusion  Regulatory element of the ERα gene was found in intron 1, but we found no association between ERα gene polymorphisms and AD.

ALZHEIMER disease (AD) is a progressive neurodegenerative disease. The APOE4 genotype is a genetic factor closely related to late-onset AD,¹ but the cause of sporadic AD has not been elucidated fully. Other genetic factors may be associated with the development of AD, and further research is essential to resolve important clinical issues and to suggest the site of the therapeutic intervention. Central cholinergic hypofunction is an established component of AD,^2,3 and estrogen replacement can significantly enhance basal forebrain cholinergic activity.⁴

Recently, estrogen replacement therapy has been reported to reduce the risk of developing AD^5-11 and to help maintain cognitive function in patients with AD.¹² The histochemical distribution of estrogen receptors (ERs) corresponds to that of cholinergic neurons, which are also capable of binding estrogen.¹³ Estrogen has been reported to increase the metabolism of amyloid precursor protein,¹⁴ improve cerebral blood flow, facilitate neuronal repair, reduce neuronal injury, and stimulate glucose transport and metabolism,¹⁵ all features that would support the use of estrogen therapy for patients with AD.

Several workers have reported associations between ERα gene polymorphisms and breast cancer,¹⁶ spontaneous abortion,¹⁷ hypertension in women,¹⁸ or osteoporosis.¹⁹ It is important to study the nontranscribed portion of this gene and search for mutations that may be involved in its regulatory domains. Restriction fragment length polymorphisms (RFLPs) of the ERα gene have been associated with a low bone mineral density in postmenopausal women.²⁰ These polymorphisms were combinations of 2 single nucleotide polymorphisms located in intron 1 and associated with AD in a Japanese population.²¹ Since these polymorphisms were located in intron 1 of the ERα gene, we speculated that RFLPs would change the enhancer activity of this gene. In an attempt to determine the contribution of this polymorphism to AD, we measured the enhancer activity of the ERα gene and then compared the activity among RFLPs. Because the enhancer activity differed among RFLPs, we particularly examined the relationship between ERα gene RFLPs and AD. We evaluated both Japanese patients clinically diagnosed as having AD and white patients diagnosed at autopsy to study ethnic variations.²²

We used the pGL3 promoter and pRL-TK vectors (Promega, Madison, Wis). The former contained the coding region of firefly (Photinus pyralis) luciferase²³ and was used as the experimental vector, whereas the latter contained renilla (Renilla reniformis) luciferase²⁴ and was used as the internal control vector. Blood samples were collected into tubes that contained an anticoagulant, and high-molecular-weight DNA was extracted from the leukocytes.²⁵ Forward (5‘-CTGCCACCCTATCTGTATCTTTTCCTATTCTCC-3‘) and reverse (5‘-TCTTTCTCTGCCACCCTGGCGTCGATTATCTGA-3‘) primers were used for polymerase chain reaction (PCR) analysis, which involved 40 amplification cycles (30-second denaturation at 94°C, 40-second annealing at 61°C, and 1-minute 30-second elongation at 72°C). A 200-ng aliquot of genomic DNA was used for each PCR solution in 50 µL containing 50-mmol/L Tris hydrochloride (pH 9.2); 14-mmol/L ammonium sulfate; 1.5-mmol/L magnesium chloride; 10% vol/vol dimethyl sulfoxide; 200-µmol/L deoxycytidine triphosphate, deoxyadenosine triphosphate, ribosylthymine triphosphate, and deoxyguanosine triphosphate; 400-nmol/L each oligonucleotide primer; and 3 U each of Taq and Pwo DNA polymerases (Boehringer Mannheim, Mannheim, Germany). The PCR product was a 1.3–kilobase pair (kbp) fragment of intron 1 and exon 2 in the ERα gene, and this fragment was digested with the restriction enzyme Pvu II or Xba I. The resulting RFLPs were coded as described previously²⁰: Pp (Pvu II) and Xx (Xba I), with uppercase and lowercase letters signifying the absence and presence of restriction sites, respectively. Each PCR product was blunt ended with T4 DNA polymerase (TaKaRa, Kyoto, Japan) and phosphorylated with polynucleotide kinase (Toyobo, Osaka, Japan). The pGL3 promoter vector was SmaI cut and dephosphorylated with calf intestinal alkaline phosphatase (TaKaRa), and the phosphorylated PCR product was inserted in both universal and reversal directions into this dephosphorylated vector using T4 DNA ligase (TaKaRa).

Transfection into HeLa S3 cells was performed using SuperFect Reagent (Qiagen, Hilden, Germany).²⁶ On the day before transfection, 6 × 10⁵ HeLa S3 cells were plated in 60-mm dishes and incubated at 37°C in an atmosphere of air that contained 5% carbon dioxide. The experimental and control vectors were double-transfected into HeLa S3 cells using 4 µg of experimental vector DNA and 400 ng of control vector DNA complexed with 8.8 µL of SuperFect Reagent. Twenty-four hours after transfection, the cells were lysed, and the luciferase assay was performed using the Dual-Luciferase Reporter Assay System (Promega). Firefly luciferase activity was measured using a luminometer, as was the renilla luciferase activity, which was used as an internal control. The firefly luciferase–renilla luciferase activity ratio represented the luciferase activity of the enhancer. The PCR products of PX, Px, and px homozygotes in both universal and reversal directions were generated and their enhancer activities compared.

All examinations were performed after obtaining informed consent from patients or their families and approving local institutional human subjects boards at all participating institutions. We collected blood samples from Japanese patients with sporadic AD, whose clinical diagnoses of probable AD were made in accordance with the criteria of the National Institute of Neurological and Communicative Disorders and Stroke–Alzheimer's Disease and Related Disorders Association,²⁷ and control blood samples from healthy Japanese volunteers. Japanese samples were from Hiroshima, Okayama, Hyogo, and Tokyo. High-molecular-weight DNA was extracted from leukocytes of these samples. White patients with AD were confirmed by neuropathologic examination at autopsy, and white controls were autopsy-examined non-AD cases.²² All white patients and controls were from Cambridge, England. Genomic DNA was extracted from the cerebral cortex of white patients with AD and control subjects using proteinase K as previously reported. In the group of Japanese patients with AD (n = 183), the mean (SD) age was 70.0 (10.2) years, and 70.5% of the patients were women; in the Japanese control group (n = 133), the mean (SD) age was 68.1 (9.1) years, and 63.9% were women; in the group of white patients with AD (n = 156), the mean (SD) age was 79.2 (10.4) years, and 55.3% were women; and in the white control group (n = 120), the mean (SD) age was 76.6 (15.4) years, and 47.5% were women. The mean age of patients with AD and control subjects was not significantly different.

The PCR analysis of subjects' DNA and restriction enzyme digestion was performed as described in the "Plasmid Construction" section. The genotype and allele distributions of patients with AD and control subjects were compared, and their APOE genes were analyzed as described previously.^28,29 Statistical analyses were performed using JMP software (SAS Institute Inc, Cary, NC), and differences at P<.05 were considered significant.

The 1.3-kbp fragment of intron 1 and exon 2 in the ERα gene was inserted into the pGL3 promoter vector. The luciferase activity in the vector with the insertion was 5.01 (n = 10, SD = 0.95) and that without the insertion was 2.15 (n = 11, SD = 0.29), which was significantly different (t test, P<.01). The weak but significant enhancer activity was verified in the 1.3-kbp fragment of intron 1 and exon 2 of the ERα gene. The luciferase activity of the different haplotypes was as follows: PX universal, 3.55 (n = 9, SD = 0.30); PX reversal, 4.06 (n = 6, SD = 0.29); Px universal, 5.54 (n = 7, SD = 0.71); Px reversal, 3.59 (n = 5, SD = 0.23); px universal, 5.65 (n = 10, SD = 0.82); and px reversal, 4.61 (n = 6, SD = 0.41). The activity of x allele seemed to be higher than X allele. The activity of the reversal types was almost the same in PX haplotype or lower than the universal types in Px and px haplotypes. The haplotype pX was not examined because pX was rare.

The genotype and ϵ4 allele frequencies were higher in patients with AD than control subjects (Table 1: Japanese genotype, χ² = 58.0, P<.001; Japanese allele, χ² = 60.7, P<.001; white genotype, χ² = 33.8, P<.001; white allele, χ² = 28.0, P<.001).^22,30 The distribution was similar to that described previously.^1,22,30 A significant difference between the APOE distribution of Japanese and white patients with AD was not observed, and no sex effect was apparent (data not shown). In Japanese patients with AD, ϵ3 allele frequency increased, and the ϵ4 allele frequency diminished with increasing age, as previously reported.³¹ However, there was no such tendency in Japanese control subjects, white patients with AD, or white control subjects (data not shown).

There was no significant difference between the ERα gene polymorphism distribution of patients with AD and control subjects (Table 2: Japanese genotype, χ² = 6.4, P = .27; Japanese PX allele, χ² = 1.4, P = .49; Japanese P allele, χ² = 1.2, P = .28; Japanese X allele, χ² = 1.0, P = .31; white genotype, χ² = 5.1, P = .74; white PX allele, χ² = 0.6, P = .91; white P allele, χ² = 0.3, P = .62; and white X allele, χ² = 0.3, P = .59).

No sex difference was detected (data not shown), but there was a racial difference. The X allele occurred more frequently in the white than the Japanese population (χ² = 13.1, P = .003). The APOE gene did not affect the ERα gene polymorphism distribution (Table 3: white, χ² = 20.9, P = .64; Japanese, χ² = 14.5, P = .49).

Although the causes of AD remain undefined, central cholinergic hypofunction is a feature of AD^2,3 and certainly contributes to the associated cognitive deficit. Loss of ovarian function had negative effects on the cholinergic neurons beyond those associated with aging and sex, suggesting that ovarian hormones play a role in maintaining the functional state of cholinergic neurons.^32,33 Recently, women were reported to be at greater risk of developing AD than men,³⁴ and estrogen replacement therapy has reduced the risk of developing AD.^5-11 Furthermore, estrogen has been shown to affect amyloid precursor protein metabolism,¹⁴ cerebral blood flow, neuronal repair,¹⁵ and synaptic sprouting.³⁵ Therefore, the possibility of a genetic link between ERα polymorphisms and a predisposition to AD²¹ seemed relevant to the pathogenesis of AD.

In this study, we detected a weak but significant enhancer activity of the 1.3-kbp fragment in intron 1 and exon 2 of the ERα gene and demonstrated that the enhancer activity differed among haplotypes. Enhancer activity of x allele seemed to be higher than X allele. The possibility that expression of ERs were regulated according to polymorphisms led us to examine the frequency of ERα gene RFLPs in patients with AD.

Although our data confirmed the association between APOE4 and AD,^1,22,30 our results did not confirm the finding that ERα gene polymorphisms were associated with AD in Japanese patients (49.4% P allele in AD vs 36.3% P allele in controls and 29.1% X allele in AD vs 16.7% X allele in controls).²¹ We failed to demonstrate this relationship in either Japanese patients or white patients with pathologically confirmed AD. The cause of these discrepant results in our Japanese patients is not clear. Sex distribution did not contribute to the different findings. The P allele frequency in control subjects was lower (36.3%) in the study by Isoe et al²¹ than in our study (40.6%) and another study²⁰ (43.6%). This may have contributed partly to the discrepancy. In contrast, the X allele distributions in control subjects did not differ among these 3 studies.

The present results were obtained from 2 different ethnic groups and with pathologically confirmed AD. A racial difference was found in the ERα gene; ie, the X allele occurred more frequently in the white than the Japanese population. Such racial differences have been reported in other genes. For example, the frequency of the A0 allele (11 dinucleotide repeats) of the tau gene was higher in Japanese than white patients,³⁶ as was the frequency of the A allele of the −491 A/T polymorphism of the APOE promoter genotype.³⁰ The APOE4 genotype was a weaker risk factor for AD among African Americans and Hispanics than among whites and Japanese.³⁷ Consequently, it is particularly necessary to study different ethnic groups to establish the relationships between polymorphisms and a disease.

In our study, the ERα gene polymorphism distribution does not differ between patients with AD and controls, whether they are Japanese or white. We do not consider, therefore, that further analysis of RFLPs in this region in relation to AD is merited.

Accepted for publication July 17, 1999.

This work was supported by grants-in-aid from the Ministry of Education, Science and Culture and from the Ministry of Health and Welfare of Japan.

We thank the Research Center of Molecular Medicine, Hiroshima University School of Medicine, for the use of their facilities and Yasuko Furuno for her technical assistance.

Reprints: Hideshi Kawakami, MD, PhD, Third Department of Internal Medicine, Hiroshima University School of Medicine, 1-2-3 Kasumi, Minami-ku, Hiroshima 734-8551, Japan (e-mail: hkawakam@ipc.hiroshima-u.ac.jp).