Glucocorticoid receptor (GR) is a transcriptional factor activated by glucocorticoids and a regulator of the expression of various genes. Human GRα (hGRα) is encoded by nine exons, including exon 9α (Encio and Detera-Wadleigh, 1991). GRβ is an alternatively spliced form with exon 9β replacing exon 9α that was identified in glucocorticoid-resistant human multiple myeloma cells and functions as an hGRα dominant negative type (Moalli et al., 1993). The 777-amino acid hGRα has many functional domains, which include DNA binding (amino acid residues 421–486), ligand binding (528–720), homodimerization (456–777), Hsp90 binding (568–653), nuclear translocation (479–506 and 526–777), and transactivation domains (77–262, 404–491, and 526–556) (Savory et al., 1999; Vottero and Chrousos, 1999). In the cytosol, GRs are associated with heat-shock and other proteins (Pratt and Toft, 1997), and the binding of glucocorticoid leads to their nuclear translocation (Webster and Cidlowski, 1999).

Glucocorticoid treatments are effective in many inflammatory diseases and some types of cancers. As for anti-inflammatory effects, GR is thought to exert its pharmacological effects through the activation of transcription, such as activating the glucocorticoid-induced leucine zipper gene, and/or through suppression of nuclear factor-κB (NF-κB) and AP-1-induced transcription of genes, such as inflammatory cytokines (Reichardt et al., 1998; De Bosscher et al., 2000; Berrebi et al., 2003). However, a small number of patients do not respond to clinically relevant doses of glucocorticoids, a condition that is termed glucocorticoid resistance (Loke et al., 2002). Several familial mutations in the GR gene have been shown to be associated with corticosteroid resistance or hematological malignancies, but these mutations are relatively rare and thought not to be a common cause of resistance (DeRijk et al., 2002). Recently, single nucleotide polymorphisms in the GR gene have been identified, some of which are relatively frequent (Panarelli et al., 1998; Ruiz et al., 2001; DeRijk et al., 2002). For example, a BclI restriction fragment polymorphism and an N363S alteration have been reported to influence the regulation of the hypothalamic-pituitary-adrenal axis and to be associated with changes in metabolism and cardiovascular control (Panarelli et al., 1998; DeRijk et al., 2002). Moreover, it has been reported that the I559N GRα variant has a transdominant effect on wild-type GR by inhibiting its nuclear translocation (Kino et al., 2002). Furthermore, the I747M GRα variant causes autosomal dominant glucocorticoid resistance through abnormal interactions with p160 steroid receptor coactivators (Vottero et al., 2002). However, little information on single nucleotide polymorphisms in the GR gene is available for the Asian population.

In this study, the coding regions of the GR gene were sequenced from 88 Japanese allergic subjects and 73 established cell lines from Japanese individuals. Two novel non-synonymous single nucleotide changes were found in addition to a previously reported adenine insertion, which induces a frameshift. Then, the effects of these nucleotide changes were assessed on the mRNA and protein expression levels, transcriptional activity, and inhibitory effect on NF-κB transactivation.

Materials and Methods

DNA Sources and Sequencing of the GR Gene. All 88 subjects used in this study were Japanese allergic subjects. Their peripheral lymphocytes were immortalized using the Epstein-Barr virus, and genomic DNA was extracted from the immortalized lymphocytes. This study was approved by both the ethical review board of the National Center for Child Health and Development and the National Institute of Health Sciences (Tokyo, Japan). Written informed consent was obtained from all participating subjects. In addition, established cell lines derived from 73 different Japanese individuals were obtained from either the Health Science Research Resources Bank (Osaka, Japan) or the Japanese Collection of Research Bioresources (JCRB), National Institute of Health Sciences (Tokyo, Japan). DNA extraction was carried out using a Blood and Cell Culture DNA kit (Qiagen GmbH, Hilden, Germany).

The primers for the polymerase chain reaction (PCR) and sequencing of the coding regions were designed based on the nucleotide sequences from the National Center for Biotechnology Information database (M78506, M78507, U78508, and AC004782). All primers used in this study have been previously described (Nagano et al., 2002). PCR conditions and sequencing methods were as previously described (Koyano et al., 2002).

Plasmids and Transfection. pRShGRα was obtained from the American Type Culture Collection (Manassas, VA), which contains the full-length coding region of human GRα and the active Rous sarcoma virus promoter (Giguere et al., 1986). Three expression plasmids encoding the variant GRs (T504S, S651F, and 2314insA) were prepared with a QuickChange Site-Directed Mutagenesis kit (Stratagene, La Jolla, CA) using the wild-type pRShGRα plasmid as a template. The empty vector plasmid used as a control was prepared by removing the hGRα cDNA from the pRShGRα plasmid. The wild-type (WT) expression plasmid (pRShGRα 0.1–0.5 μg) was first titrated with 0.6 μg of the mouse mammary tumor virus (MMTV) promoter-luciferase reporter construct (pHH-Luc, the reporter plasmid described under Luciferase Reporter Assay) using COS-7 cells (1.4 × 10⁵ cells) for Western blot analysis and luciferase activity. These results indicated that protein expression levels and luciferase activities linearly increased depending on the amounts of the transfected plasmid (data not shown). Thus, 0.3 μg of the WT or variant expression plasmids per 1.4 × 10⁵ cells was utilized for transfection. phRL-TK (Promega, Madison, WI) encoding Renilla (sea pansy) luciferase was used for cotransfection and normalization of transfection efficiency. To assess the efficiency of each transfection, the Renilla luciferase activity was measured from transfected cell lysates. All of the transfection efficiencies were consistent in Northern and Western blot analyses (92–103% of the WT hGRα expression plasmid among all of the variant expression plasmids). The PathDetect NF-κB cis-Reporting System (Stratagene) was used in the inhibition experiments for NF-κB transcriptional activity by the GR variants. pFC-MEKK (Stratagene) expressing constitutive active MEKK (amino acid 360-672) was used as a positive control activating NF-κB. pNF-κB-luc (Stratagene) was a 5xNF-κB binding element-driven firefly luciferase plasmid.

Cell Culture. COS-7, an African green monkey kidney cell line, was chosen for transfection of the GR expression plasmids since the endogenous GR levels are low (see Figs. 1 and 2). COS-7 cells were obtained from JCRB and grown in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum and 100 U/ml penicillin and streptomycin under an atmosphere of 5% CO₂ at 37°C. For dexamethasone treatment, cells were cultured in Dulbecco's modified Eagle's medium supplemented with fetal bovine serum (10%) that was pretreated with dextran-coated charcoal to remove endogenous glucocorticoids.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

Northern blot analysis of the WT, T504S, S651F, and 2314insA variant GR transcripts. Total RNA samples were prepared from COS-7 cells transfected with 3.5 μg of the empty vector, WT, or variant GR expression plasmids, together with 0.5 μg of the phRL-TK plasmid. Five micrograms of the RNA samples were separated by electrophoresis, transferred to a nylon membrane, and hybridized with ³²P-labeled GR cDNA probe as described under Materials and Methods (upper panel). The applied RNA quantity was confirmed by endogenous GAPDH hybridization signals shown in the lower panel. One of the representative results from three independent experiments is shown.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

Western blot analysis of the WT and variant GR proteins. Whole cell lysates were prepared from COS-7 cells, which were transfected with 3.5 μg of the vector, WT, or each variant GR expression plasmid, together with 0.5 μg of the phRL-TK plasmid. The lysates were subjected to electrophoresis and blotted as described under Materials and Methods. One of the representative results from three independent experiments is shown.

Northern Blot Analysis. COS-7 cells (1.6 × 10⁶ cells) were cotransfected with 3.5 μg of expression plasmid for the wild-type or variant GRs and 0.5 μg of the phRL-TK plasmid using the PolyFect Transfection Reagent (Qiagen). The cells were harvested 48 h after transfection, and total RNA was extracted with the RNeasy kit (Qiagen). Because all of the transfection efficiencies were similar, as described above, an equal amount of total RNA (5 μg) from each RNA sample was subjected to electrophoresis, transferred to a nylon membrane (GeneScreen Plus; PerkinElmer Life Sciences, Boston, MA), and hybridized with ³²P-labeled hGRα or monkey glyceraldehyde-3-phosphate dehydrogenase (GAPDH) cDNA probe as previously described (Brown and Mackey, 1997). hGRα cDNA was generated by digesting the pRShGRα plasmid with KpnI and XhoI. Monkey GAPDH cDNA was prepared by reverse transcription-PCR using monkey GAPDH-specific primer pairs and COS-7 cell total RNA as a template. The hybridized signals were analyzed with a Bioimage Analyzer BAS-1500 (Fujifilm, Tokyo, Japan). The expression levels were quantified from three separate transfection experiments.

Western Blot Analysis. The cotransfected COS-7 cells (1.6 × 10⁶ cells) were prepared as described above, and the cell pellets were boiled in a protein sample buffer. Because all of the transfection efficiencies were similar, as described above, an equal amount of total protein (20 μg) was separated on a 6% SDS-polyacrylamide gel (Tefuco Co., Tokyo, Japan) and transferred onto a polyvinylidene difluoride membrane (Bio-Rad, Hercules, CA). The membrane was blocked with 5% skim milk and incubated with polyclonal rabbit anti-human GR antibody (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) and then with goat anti-rabbit IgG conjugated with horse-radish peroxidase (Amersham Biosciences UK, Ltd., Little Chalfont, Buckinghamshire, UK) for the secondary antibody. GR protein was visualized with the SuperSignal West Femto maximum sensitivity substrate (Pierce Chemical, Rockford, IL) according to the manufacturer's instructions. The signals were assessed by densitometric analysis. The expression levels were quantified from three separate transfection experiments.

Time Course Experiments of 2314insA mRNA and S651F and 2314insA Proteins. Time course experiments were performed essentially as previously described (Webster et al., 1997). For the measurement of mRNA levels, COS-7 cells were transfected with GR WT or the 2314insA expression plasmid as described above and 32 h later were treated with 1 μg/ml of actinomycin D for 1 h, followed by a 24-h incubation. Total RNA was prepared after 0, 3, 6, 12, and 24 h and analyzed by Northern blotting. Two independent transfection experiments were performed. For protein level measurements, COS-7 cells were transfected with the GR WT, S651F, or 2314insA expression plasmid and 42 h later were treated with 1 μM cycloheximide for 1 h, then incubated for an additional 48 h. Whole cell lysates were prepared after 0, 3, 6, 12, 24, and 48 h and subjected to Western blotting. The expression levels were quantified from three separate transfection experiments.

Immunocytochemistry. COS-7 cells (1.6 × 10⁶ cells) were transfected with 4 μg of the expression plasmid for wild-type or variant GR and then cultured for 48 h. The cells were treated with vehicle or 100 nM dexamethasone for 90 min. Transfected COS-7 cells were washed twice with PBS, fixed with 3.7% formaldehyde in PBS for 15 min at room temperature, and then permeabilized with 0.5% Triton X-100 for 15 min. The fixed cells were blocked by incubation with PBS containing 10% goat serum and simultaneously treated with 100 ng/ml of DAPI (Santa Cruz Biotechnology, Inc.) for 30 min at room temperature and then washed with PBS containing 0.05% Tween 20. Next, the cells were incubated overnight at 4°C with rabbit polyclonal anti-GR antibody (Santa Cruz Biotechnology, Inc.) diluted in PBS containing 10% goat serum. They were washed five times with PBS containing 0.05% Tween 20 and incubated for 1 h at room temperature with anti-IgG (rabbit) conjugated with Alex 594 (Funakoshi Co., Tokyo, Japan) diluted in PBS containing 10% goat serum. Immunoreacted GR and the nuclei stained with DAPI were visualized by fluorescent microscopy.

Luciferase Reporter Assay. COS-7 cells (1.4 × 10⁵ cells) were cotransfected with 0.3 μg of the wild-type GR expression plasmid pRShGRα or the variant GR expression plasmid together with 0.6 μg of MMTV promoter-luciferase reporter construct pHH-Luc (American Type Culture Collection) (GenBank accession number AF093686) and 0.6 μg of phRL-TK plasmid as an internal control. Twenty-four hours after transfection, the cells were treated with vehicle (methanol) or various concentrations of dexamethasone (DEX), then cultured for an additional 24 h. The cells were washed with PBS, and the lysates were prepared using the PicaGene Dual Sea Pansy Luminescence kit (Nippon Gene Co., Ltd., Tokyo, Japan). Luciferase activity was then measured using a VICTOR² Multilabel Counter (PerkinElmer Wallac, Turku, Finland). All transfection efficiencies were normalized according to the Renilla luciferase activity. Quantification was done from three independent transfection experiments.

Inhibition of NF-κB Transcriptional Activity by the GR Variants. The Pathdetect NF-κB cis-reporting system (Stratagene) was used in the assay for inhibition of NF-κB transcriptional activity of the GR variants. COS-7 cells (2.8 × 10⁴ cells) were cotransfected with 5 ng of 5xNF-κB-luc reporter plasmid and 50 ng of phRL-TK, with or without 50 ng of either the WT or variant (T504S, S651F, or 2314insA) expression plasmids, and 5 ng of pFC-MEKK. Twenty-four hours after transfection, the cells were treated with 100 nM DEX for an additional 43 h. The cell lysates were prepared, and their luciferase activities were measured. All transfection efficiencies were normalized according to Renilla luciferase activity. Quantification was done from three independent transfection experiments.

Statistical Analysis. The results of the Northern blot analysis, Western blot analysis, and reporter assays were assessed for their statistical significance by one-way analysis of variance, followed by Fisher's PLSD method using StatView software (SAS Institute Inc., Cary, NC).

Results

Three Genetic Variations Found in the GR Gene. The coding regions of the GR gene were sequenced from 88 Japanese subjects and 73 established cell lines from Japanese individuals. From the 88 Japanese subjects, two novel non-synonymous single nucleotide alterations, which are substitutions at nucleotide 1510 (1510A>T; T504S) (based on the reference sequence; GenBank accession number AC004782) and at nucleotide 1952 (1952C>T; S651F), were found. In addition, an insertion of an adenine between nucleotides 2313 and 2314 (2314insA) was identified. All three variants were heterozygous at an allelic frequency of 0.0031. The insertion 2314insA generates a frameshift from codon Leu772 (exon 9α), resulting in a substitution of six amino acid residues and an additional 20 residues at the end of the C terminus (LLFHQK to TSVSSKVTALIRMVALKKVELIAFIV). This insertion was detected as a heterozygote in cell line CCK-81 (Health Science Research Resources Bank accession number JCRB0208), which was derived from a Japanese patient with colon adenoma (Koura and Isaka, 1980; Sakurai et al., 1988) and is identical to the insertion that Jiang et al. (2001) found in 8 of 39 lupus nephritis patients. We also characterized the functional effect of this insertion, since its functional significance was unclear.

Expression of the Variant GR mRNA in the Transfected COS-7 Cells. To determine expression levels of the variant GR mRNAs, Northern blot analysis was performed using total RNA from COS-7 cells transfected with the wild-type or variant expression plasmids (n = 3) (Fig. 1). From Northern blot analysis, the GR mRNA expression levels of the T504S, S651F, and 2314insA variants were estimated at 113.7 ± 12.2, 85.9 ± 5.5, and 35.6 ± 1.0 (p < 0.005), respectively, when the expression level of the WT GR was defined as 100. Thus, the mRNA expression levels of T504S GR and S651F GR were similar to that of the WT, but the 2314insA GR expression was significantly reduced.

Expression Levels of Variant GR Proteins in Total Cell Lysates. To quantify the expression levels of the variant GR proteins, we performed Western blot analysis of total cell lysates from the transfected COS-7 cells using an anti-GR antibody (n = 3). By densitometric quantification, the protein expression levels of the T504S, S651F, and 2314insA GR variants were estimated at 90.2 ± 15.1, 65.6 ± 4.3 (p < 0.05), and 6.3 ± 1.7 (p < 0.001), respectively, when the WT expression level was defined as 100. No significant difference in the expressed protein levels was found between the WT and T504S (Fig. 2). However, S651F expression was significantly reduced, and the 2314insA protein levels were drastically reduced (Fig. 2). Considering the results of the Northern blot analysis, it is likely that the reductions seen in the S651F and 2314insA protein levels were caused by protein instability or ineffective translation in COS-7 cells in addition to mRNA instability for 2314insA.

GR Variant mRNA and Protein Time Course Experiments. Because the GR variants showed reduced expression of mRNA (2314insA) and protein (S651F and 2314insA) compared with the WT, as shown in Figs. 1 and 2, time course experiments were performed for the 2314insA mRNA and the S651F and 2314insA proteins. For mRNA, the amount of 2314insA mRNA (t_1/2 = ∼9 h) decreased faster than that of WT mRNA (t_1/2 = ∼20 h) in COS-7 cells in the presence of actinomycin D (n = 2) (Fig. 3A). For protein, half-lives of the WT, S651F, and 2314insA proteins in the presence of cycloheximide were approximately 12, 9, and 6 h, respectively (n = 3) (Fig. 3B). These results suggested that instabilities of the 2314insA mRNA and protein and of the S651F protein cause their reduced protein levels.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

Time course experiments of 2314insA mRNA and S651F and 2314insA proteins. COS-7 cells were transfected with either WT, S651F, or 2314insA expression plasmids. Thirty-two (for mRNA) or forty-two (for protein) hours after transfection, the cells were treated with 1 μg/ml of actinomycin D for mRNA stability or 1 μM of cycloheximide for protein stability for 1 h, then total RNA (WT and 2314insA) or total cellular protein (WT, S651F, and 2314insA) was prepared after the indicated periods of time (h), and the expression levels were assessed by Northern or Western blotting. The Western blot data were quantified by densitometric analysis. A, Northern blots for WT and 2314insA. One of the representative results from two independent experiments is shown. B, protein expression levels for WT, S651F, and 2314insA. Data from three independent experiments are shown. Bars represent the mean ± S.D. (n = 3).

Subcellular Localization of the Variant GR. GR exists in the cytoplasm in a glucocorticoid-free form (Sanchez et al., 1986; Gustafsson et al., 1989). To investigate the subcellular localization of each variant GR, immunostaining of COS-7 cells transfected with the WT or variant GR plasmids was done (n = 3) (Fig. 4). When WT-transfected cells were not treated with DEX, the expressed receptors were cytoplasmic. In contrast, when cells were treated with 100 nM DEX, most of the wild-type GR proteins were localized in the nucleus. The localization patterns of T504S and S651F variant proteins were similar to that of the WT (Fig. 4). In contrast, the 2314insA variant protein was barely detectable in the cytoplasm (data not shown), which was consistent with the Western blot results. Long exposure showed that the 2314insA protein was distributed in the cytoplasm even if the cells were treated with 100 nM DEX (Fig. 4). Thus, this variant protein is suggested to also lack glucocorticoid-dependent nuclear translocation ability.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 4.

Cellular localization of hGRα and its variant proteins in COS-7 cells. COS-7 cells were transfected with the WT hGRα (WT) or each variant (T504S, S651F, and 2314insA) expression plasmid in the absence (–; left two columns) or presence (+; right two columns) of 100 nM DEX. The transfected cells were immunostained with anti-hGR antibody and Alex 594-conjugated second antibody. Photographs in the left column show the immunoreacted WT and variant proteins (red). Photographs in the right column show DAPI-stained nuclei (blue). The empty plasmid-transfected cells (negative control) were also immunostained (empty). Data from one of three independent transfections are shown. In the case of 2314insA (red), the exposure time was ca. 20-fold longer than the other variants.

Transcriptional Activity of the Variant GRs. GRs induce expression of a variety of genes. GRs are thought to also exert their anti-inflammatory effects partly via activation of transcription of the genes, such as recently identified glucocorticoid-induced leucine zipper genes, the proteins of which interfere with NF-κB and AP-1 function (Mittelstadt and Ashwell, 2001; Berrebi et al., 2003). To investigate the transcriptional activity of the variant GRs, COS-7 cells were transfected with the WT or variant GR expression plasmid together with a plasmid containing the MMTV promoterluciferase reporter construct (pHH-Luc). The MMTV promoter has been extensively used as a model system to study the mechanisms of glucocorticoid-mediated gene regulation (Medh and Schmidt, 1997) and contains a well characterized glucocorticoid-responsive element that mediates transcriptional activation. The results of the reporter gene experiments are shown in Fig. 5 (n = 3). Transfection of COS-7 cells with the empty vector plasmid alone induced no luciferase activity, even with dexamethasone treatment. When the WT was transfected, luciferase activity increased in a dose-dependent manner with dexamethasone. Similarly, the T504S variant showed an increase in luciferase activity comparable with that of the WT. In contrast, the S651F variant treated with 100 nM dexamethasone showed a significant decrease to approximately 62.9 ± 13.3% (p < 0.001) of the WT levels. Furthermore, the 2314insA variant GR induced little or no luciferase activity. The transcriptional activities of the variant GRs correlated well with their protein expression levels (Fig. 2).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 5.

Transcriptional activity of the WT and variant GRs. COS-7 cells were transfected with an expression plasmid, pHH-Luc reporter plasmid, and phRL-TK plasmid. Twenty-four hours after transfection, the cells were treated with vehicle or various concentrations of DEX as indicated and continuously cultured for an additional 24 h. Luciferase activity is shown as the percentage activity of WT with 100 nM DEX. Bars represent the mean ± S.D. (n = 3). **, p < 0.001 by Fisher's PLSD method.

Inhibition of NF-κB Activation by the GR Variants. It has been generally thought that the GR suppresses inflammation by transcription inhibition of proinflammatory genes induced by other nuclear factors, such as NF-κB (Reichardt et al., 1998; De Bosscher et al., 2000). Thus, we examined the inhibitory ability of the GR variants on NF-κB transactivation (n = 3) (Fig. 6). NF-κB transactivation was assayed using the 5xNF-κB enhancer-containing luciferase plasmid and the MEKK expression plasmid. This assay system used endogenous NF-κB in COS-7 cells (Reinhard et al., 1997). The 5xNF-κB-driven luciferase activity increased more than 200-fold with cotransfection of the constitutively active MEKK expression plasmid as compared with that of the cotransfected vehicle alone (data not shown). By cotransfecting the WT expression plasmid, the NF-κB-dependent 5xNF-κB-driven luciferase activity was reduced to 17.9 ± 3.8% (p < 0.001) compared with cotransfection of the MEKK expression plasmid alone (data not shown). When inhibitory effects of WT were defined as 100%, the NF-κB-dependent luciferase activities were inhibited to 100 ± 5.2%, 81 ± 9.4% (p < 0.005), and 12 ± 2.9% (p < 0.001) when cotransfected with T504S, S651F, and 2314insA expression plasmids, respectively (Fig. 6). These results indicated that the inhibitory effect of the GR variants on the NF-κB transactivation almost correlates with their protein expression levels, similar to their transcriptional activation properties.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 6.

Inhibitory activity of the GR variants on NF-κB activation. COS-7 cells were transfected with an expression plasmid encoding WT, T504S, S651F, or 2314insA together with pNF-κB-luc, pFC-MEKK, and phRL-TK as described under Materials and Methods. Twenty-four hours after transfection, the cells were treated with 100 nM of DEX and cultured continuously for an additional 43 h. All firefly luciferase activities were normalized by Renilla luciferase activities. Bars represent the mean ± S.D. (n = 3). *, p < 0.005 and **, p < 0.001 by Fisher's PLSD method.

Discussion

GR is a transcriptional regulatory protein, directly interacting with DNA, and also is a modulator of transcription mediated by other transcription factors. To identify the single nucleotide polymorphisms of the GR gene affecting the glucocorticoid responses, the coding regions of the GR gene were sequenced. Polymorphisms at 1510A>T (T504S) and 1952C>T (S651F) were found in Japanese subjects, and an adenine insertion between nucleotides 2313 and 2314 (2314insA) was identified in a cell line. Then, the effects of the nucleotide variations on mRNA (Fig. 1), protein levels (Fig. 2), and glucocorticoid-responsive element-mediated transcriptional activity (Fig. 5) were assessed.

To date, it has been shown that NF-κB and GR have opposing actions in the modulation of the immune/inflammatory responses (McKay and Cidlowski, 2000). NF-κB induces the expression of proinflammatory genes, whereas GR suppresses immune function, in part, by reducing the expression of the same genes. The molecular mechanism of GR function is still controversial, but its direct binding to the NF-κB p65 subunit is thought to be involved in the subsequent suppression of NF-κB-mediated transactivation (Reichardt et al., 1998; De Bosscher et al., 2000; Nissen and Yamamoto, 2000). Therefore, we also examined an inhibitory effect of the GR variants on NF-κB transcriptional activation (Fig. 6). In addition, GR also transactivates several genes including the recently identified glucocorticoid-induced leucine zipper gene, the protein of which interferes with the function of NF-κB and AP-1. Thus, glucocorticoid-induced transactivation is also suggested to be involved in its anti-inflammatory effects (Mittelstadt and Ashwell, 2001; Berrebi et al., 2003).

Recently, Jiang et al. (2001) reported the same adenine insertion polymorphism (insertion of A at 2439) in 8 of 39 lupus nephritis patients and claimed that the lupus nephritis patients had a lower glucocorticoid-binding capacity in peripheral blood mononuclear cells. Thus, although we detected this insertional variation in an established cell line, this is a true polymorphism, not a mutation formed during tumorigenesis. Because they examined the ligand-binding capacity for mixed samples of the variant and wild-type GRs, the functional changes caused by the insertion were unclear. Thus, our study determined, for the first time, the functional changes caused by this insertion.

For the 2314insA variant, the mRNA level, protein level, transcriptional activity, and inhibitory effects on NF-κB were reduced to approximately 36, 6, 2, and 12% of the WT, respectively (Figs. 1, 2, 5, and 6). This indicated that the insertion influences both mRNA and protein expression, and its reduced transcriptional activity and inhibitory effects on NF-κB almost correlated with its protein level. Because the variant expression plasmids had the same Rous sarcoma virus promoter region, it was likely that the reduced mRNA levels in the 2314insA variant were caused by mRNA instability in the COS-7 cells. This was confirmed by the time course experiment (Fig. 3A), which showed that the expression of the 2314insA mRNA decreased faster than that of the GR WT mRNA (Fig. 3A). Nonsense-mediated mRNA decay has been reported in many genes to destroy mRNA species that contain premature termination codons (Byers, 2002). Our case is different from the nonsense-mediated mRNA decay, since the 2314insA frameshift variant creates a stop codon 52 base pairs downstream from the normal one. However, Simpson and Stoltzfus (1994) have reported a v-src mutant mRNA, which causes a frameshift, resulting in the formation of a novel termination codon 13 base pairs downstream from the normal termination site of the src polypeptide. In this case, the expression level of the v-src mutant mRNA was reduced by approximately 40 to 50% of that of the wild type. It is possible that the mechanism of the reduced mRNA expression of the 2314insA variant is similar to the reduced expression of the v-src mutant mRNA. Since the half-life of the 2314insA protein was also reduced as compared with the WT (6 versus 12 h) (Fig. 3B), a significantly lower expression of the 2314insA protein was caused by both its mRNA and protein instabilities. This variant protein also suggested that it lacked glucocorticoid-dependent nuclear translocation ability (Fig. 4). Furthermore, the 2314insA variant lacks the functional activation function-2 domain, which is similar to hGRβ. Therefore, we have examined whether the 2314insA variant acts as a dominant negative type, like GRβ. COS-7 cells were transfected with the pHH-Luc reporter plasmid and equal amounts of 2314insA variant and/or WT expression plasmids and cultured in the presence of 100 nM dexamethasone. The luciferase activity of the COS-7 cells coexpressing 2314insA and WT was almost equal to that expressing WT alone, suggesting that 2314insA did not act as a dominant negative type against the WT GR (data not shown). Similar results were obtained for the inhibition of NF-κB-mediated transactivation by GR (data not shown).

The S651F variant was found in an atopic dermatitis patient with a relatively poor response to glucocorticoid drugs. The protein level, transcriptional activity, and inhibitory effect on NF-κB of S651F were reduced to approximately 66, 63, and 81%, respectively, compared with the wild type, without any changes in mRNA levels (Figs. 2, 5, and 6). This suggested that the reduction in the protein levels was attributed to protein instability or translational inefficiency. Then, the S651F protein half-life in COS-7 cells was determined. Time course experiments showed that the half-lives of WT and S651F were about 12 and 9h, respectively (Fig. 3B). Because the S651F mRNA was stable, similar to the WT mRNA (Fig. 1), reduced expression of the S651F protein was due only to its protein instability. Residue 651 is located in the Hsp90-binding domain. It is possible that the replacement of Ser-651 with phenylalanine affects the Hsp90 binding to the variant GR, since Hsp90 stabilizes the GR protein. The S651F variant was able to translocate to the nucleus in response to dexamethasone (Fig. 4).

Although Thr-504 is located in the nuclear translocation domain, the T504S protein translocates into the nucleus in response to dexamethasone (Fig. 4). This T504S variation was found in an atopic dermatitis patient who had a good response to glucocorticoid drugs. The T504S variant showed similar levels of mRNA, protein, transcriptional activity, and inhibitory effect on the NF-κB transactivation compared with those of the wild type.

The results in this study demonstrated that the S651F and 2314insA variants have partially reduced and almost abrogated activities for transcription and NF-κB inhibition, respectively, almost paralleling their reduced protein expression levels. Thus, these two variations may influence the response to glucocorticoid treatment.