| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
* Nephrology and Hypertension and Clinical Research;
Institute for Clinical Chemistry, University Hospital of Berne, Berne, Switzerland;
Canada Research Chair for Cardiovascular Obesity Research and Management, McMaster University, Hamilton General Hospital, Hamilton, Ontario, Canada; and
Department of Psychosomatics and Psychotherapy, Charité Campus Benjamin Franklin, Berlin
1Correspondence: Department of Nephrology and Hypertension, University Hospital, Freiburgstrasse 15, 3010 Bern-Inselspital, Switzerland. E-mail: brigitte.frey{at}dkf.unibe.ch
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: hypertension • NF1 • glucocorticoid receptor • glucocorticoid responsive element • microsatellite
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 2)
. Functional assays of the ancestral corticoid receptor revealed activation by very low doses of aldosterone and cortisol, the physiological ligands of the MR and GR in humans (1)
. Given the common phylogenetic origin of these two receptors, the absence of ligand-specificity is not a surprise. For instance, cortisol exhibits a similar affinity for the MR as does cortisol (3
4
5)
. Therefore, aldosterone can only act specifically as the dominant MR activator whenever cortisol is inactivated in MR-expressing cells. This inactivation of cortisol is performed by the enzyme 11beta-hydroxysteroid dehydrogenase (11beta-HSD2; refs. 3
4
5
).
The best-established physiological role for the 11beta-HSD2 enzyme is the regulation of salt homeostasis by preventing access of glucocorticoids to the MR in sodium transporting epithelia (3
4
5)
. When 11beta-HSD2 is impaired, cortisol acts as a mineralocorticoid. A potentially fatal genetic disorder, the syndrome of "apparent mineralocorticoid excess," has been attributed to mutations in the HSD11B2 gene (5
, 6)
. The phenotype is characterized by increased tubular sodium retention with low renin, low aldosterone hypertension. A moderate, but clinically significant, decrease in the activity of 11beta-HSD2 can be attributable either to 1) endogenous or exogenous enzyme inhibitors (3
, 5
, 7
, 8)
, 2) a mutated protein with partially reduced activity (6
, 9)
, or, possibly, 3) a decreased enzyme expression due to polymorphisms in regulatory regions. So far no polymorphisms in regulatory regions of the HSD11B2 gene have been identified, although HSD11B2 is a promising candidate for such variants for the following reason. Several studies (10
11
12
13
14
15)
suggest that reduced 11beta-HSD2 activity may play a role in essential hypertension and in sensitivity of blood pressure to dietary salt. Given the low prevalence of nonconservative mutations in the coding-region in this gene (<1/250,000; ref. 16
), the reduced 11beta-HSD2 activity observed in these studies was not explained by mutations resulting in altered amino acid sequences. An abnormal regulation of the HSD11B2 gene expression is more likely than nonconservative exonic mutations for determining the activity of 11beta-HSD2. Therefore, we and others (17
18
19
20
21
22)
have investigated the transcription factors responsible for regulation of 11beta-HSD2 expression. So far Sp1 and Sp3 are relevant for basal expression, whereas Egr1 and NF-
B (p50) inhibit and NF-B (p65) and NF1 enhance the HSD11B2 expression. CpG methylation of promoter regions reduces the affinity of activating factors including NF1 and Sp1/Sp3 and enhances binding of the MBD2 and MBD1 repressor complexes to the promoter (17)
.
Recent studies (10
, 12
, 13
, 23
24
25)
have shown that there is an association between an HSD11B2 flanking microsatellite and hypertension. Thus, we were interested in analyzing the HSD11B2 promoter for possible polymorphisms and in correlating polymorphisms and the urinary ratios of cortisol to cortisone metabolites to microsatellite pattern. We identified four novel variants. Molecular analyses revealed that two common variants (G-209A and G-126A) have an important effect on HSD11B2 promoter activity. The proportion of individuals carrying these two polymorphisms was elevated in subjects with salt sensitivity. Additionally, the polymorphism G-209A was associated with the allele pair A4/A7.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Subjects
Genetic analysis of the HSD11B2 promoter was investigated in 96 healthy male, salt-resistant and 57 healthy, salt-sensitive subjects (6 females, 51 males; refs. 10
, 26
, 27
). Ages ranged from 20–30 years. Physical findings, blood pressure, and urinary electrolytes are published elsewhere (10
, 23)
. Salt- sensitive subjects (SS) were compared to a similar salt-insensitive collective of healthy volunteers (SR). To rule out hypertension, hyperlipidemia, diabetes mellitus, hepatic, or renal diseases, volunteers were examined using physical and laboratory tests. Only subjects with a diastolic blood pressure below 85 mmHg and a systolic pressure below 140 mmHg were included. Parental histories of hypertension were communicated personally by family physicians. Subjects with one parent treated for hypertension were excluded. All participants gave informed consent, as required by the ethical committee of our institutions.
Polymerase chain reaction analysis
Genomic DNA mutation/polymorphism in the HSD11B2 promoter was detected by single-strand conformational polymorphism (SSCP) of polymerase chain reaction (PCR) products from the proximal promoter and part of exon-1 (forward 5'-GTGGCATGTGCTCACCTGAGCG-3'; reverse 5'-AGACGCAGGTCTGAGCGCAGCA-3' primers). Reactions were performed with 0.4 µM of each primer in a buffer containing 2 mM MgCl2 and 0.5 mM of each dNTP using the GC-RICH PCR System (Roche). DNA was amplified for 35 cycles, denatured at 94°C, 20 s, annealed at 63°C, 30 s, and extended at 72°C, 45 s. PCR products were analyzed on 12% acrylamide gels and visualized by silver staining. Double-band shifts indicated sequence changes. Purified variants (QIAquick, Qiagen) were sequenced (Microsynth). Two variants were analyzed for restriction sites.
PCR analysis and sequencing of the polymorphic microsatellite marker were performed by the automated fluorescent genotyping method using primers and instruments described previously (10)
. For this purpose, only SS (n=54) and SR (n=96) were investigated, of which urinary steroid profiles of 24 h urine collections were available.
Urinary steroid profile analysis
Steroid profiles of 24 h urine collections were analyzed in our laboratory by gas chromatography-mass spectrometry (GC-MS) according to published protocols (7)
.
Cell cultures
Human adenocarcinoma cells SW620 (colon), MCF-7 (breast), and Ishikawa (endometrium) were grown in FCS supplemented DMEM. To study up-regulation of 11beta-HSD2, cells were exposed to 0.1 µM dexamethasone or 1 µM RU38486 in serum-free medium. Levels of mRNA and activity were determined.
Reverse transcription
Reverse transcription (RT) was performed in RT buffer containing 50 mM Tris-HCl (pH 8.2), 6 mM MgCl2, 10 mM DTT, 100 mM NaCl, 250 nM dNTPs, 2 U of ribonuclease inhibitor RNAsin, 1 U avian myeloblastosis virus reverse transcriptase, 2 µl of random hexanucleotide mix (10-fold concentrated), and 2 µg of total RNA in a volume of 20 µl for 1 h at 42°C.
Quantitative PCR
PCR was performed using the TaqMan system from Applied Biosystems (Foster City, CA, USA). The reaction mixture contained TaqMan universal PCR master mix, 500 nM each of the appropriate forward and reverse primers, 200 nM of specific TaqMan probe, and 100 ng of reverse transcribed total RNA in a final volume of 25 µl. Primers and probe were designed with the Primer Express 1.0 software (Applied Biosystems), optimized, and validated. The primers were complementary to the human 11beta-HSD2 cDNA (forward primer position 802–821, reverse 850–869, probe 823–847). TaqMan glyceraldehyde-3-phosphate dehydrogenase (GAPDH) RNA control reagents were used for normalization (20-fold concentrate, Applied Biosystems). Probes were labeled with a reporter fluorescent dye, 6-carboxyfluorescein (FAM), and with 6-carboxy-tetramethyl-rhodamine (TAMRA) as quencher. The prerun cycling conditions were 2 min at 50°C and 10 min at 95°C. Thermal cycling involved 40 cycles at 95°C for 15 s and 1 min at 60°C. Primers and probe of human 11beta-HSD2 (forward position; 802–821, reverse: 850–869, probe: 823–847) were designed with Primer Express (Applied Biosystems). GAPDH was used for normalization. Prerun cycling conditions were 2 min, 50°C and 10 min, 95°C. Thermal cycling involved 40 cycles at 95°C, 15 s and at 60°C, 1 min.
11beta-HSD2 activity assay
Cells were scraped and lysed in sucrose buffer (250 mmol/l sucrose, 2 mmol/l EDTA, and 20 mmol/l Tris-HCl, pH 7.5), and the supernatant was stored at –20°C. Protein content was determined using the Bradford protein assay (Bio-Rad, Glattbrugg, Switzerland). The cortisol/cortisone conversion was used to measure oxidation at C-11 by 11beta-HSD2 following protocols described earlier (28)
. In brief, cell homogenates were incubated with 5 nCi [3H]cortisol, 10 nmol/l cortisol, and 200 µmol/l nicotinamide adenine dinucleotide in sucrose buffer. The reaction was stopped by adding 20 µl of 1mg/ml unlabeled cortisol and cortisone in methanol and TLC developed in chloroform-methanol (90:10 v/v). Steroids were located using ultraviolet light, excised, and counted in a Packard scintillation counter (Tri-Carb 2000CA; United Technologies, Hartford, CT, USA). Specific activity was expressed as picomoles per micrograms protein per hour. Control samples from each cell line were incubated at the same condition in the presence of Triton X-100 to abrogate 11beta-HSD2 activity and to measure 11beta-HSD1 activity (29)
. The assay was repeated up to four times using different protein concentrations within each individual experiment.
Plasmid constructs
Reporter plasmid pGL3–400 and pGL3-XhoI containing part of promoter and exon 1 (pGL3–400: –375 to +132 and pGL3-Xho: –196 to +132, based on kidney transcription start) of human HSD11B2 gene promoter were cloned into pGL3-Basic. Mutant plasmids Mut-(–209), Mut-(–126), Mut-(–194), Mut-(–151), and pGL3–400(GRE-) and pGL3-Xho(GRE-) derivatives of plasmids pGl3–400 or pGL3-XhoI were generated via PCR-based whole plasmid synthesis using Pyrococcus furiosus(Pfu) polymerase. Primers were for Mut-(–209): forward: 5'-CCAAGCACCGCCCGCAACCAGGCGGCTCCTCG-3', reverse: 5'-CGAGGAGCCGCCTGGTTGCGGGCGGTGCTTGG-3'; for Mut-(–194): forward: 5'-CCAGGCGGCTCCTCCAGCGCAGCAACTTTGG-3', reverse: 5'-CCAAAGTTGCTGCGCTGGAGGAGCCGCCTGG-3'; for Mut-(–151): forward: 5'-CCGGCTTTTTCCAAATCAAATCTGGTCGAGGGGGCGG-3', reverse: 5'-CCGCCCCCTCGACCAGATTTGATTTGGAAAAAGCCGG-3'; for Mut-(–126): forward: 5'-CGAGGGGGCGGGGCGGAGGGGGAGCACCTGC-3', reverse: 5'-GCAGGTGCTCCCCCTCCGCCCCGCCCCCTCG-3'; for pGL3–400(GRE-) and pGL3-Xho(GRE-): forward: 5'-GCAGCAACTTTGGGACTTTCATCCGGCTTTTTCCAAATCG-3', reverse: 5'-CGATTTGGAAAAAGCCGGATGAAAGTCCCAAAGTTGCTGC-3'. Parental templates were digested with DpnI. Murine NF1 and rat GR plasmids were gifts from Dr. Gronostajski (30)
and Dr. Rusconi (31)
.
Transfection and reporter gene assay
DNA (QIAfilter columns) was transfected with FuGene6 (Roche) in 24-well plates. Each transfection consisted of an experimental construct (0.7 µg), Renilla luciferase–thymidine kinase (RL-TK) (Promega), internal control (0.11 µg), and 0.25 µg of various NF1 expression constructs. Where indicated, 0.25 µg of rat GR were cotransfected. Carrier DNA (CMV-vector) was added to a total of 1.31 µg/well. After 12–16 h, washed cells were incubated for 24 h in serum-free medium with or without 0.1 µM dexamethasone. Cells were harvested in 100 µl/well reporter lysis buffer (Promega), and chemiluminescence of Firefly and Renilla luciferases was measured with MediatorsPhL luminometer (Mediators Diagnostic Systems, Vienna, Austria). Transfections were confirmed by multiple independent experiments.
Electrophoretic mobility shift assay
Nuclear extracts (NE) were prepared (21
, 22)
. Top strand sequences of synthetic complementary oligonucleotides are indicated in Table 1
. Nuclear extracts (3–4 µg) were incubated with 35 fmol of 
[32P]-ATP labeled oligonucleotides for 20 min at room temperature in binding buffer (4% Ficoll, 20 mM HEPES, pH 7.5, 35 mM NaCl, 60 mM KCl, 0.01% Nonidet P-40, 2 mM DTT, and 0.7–1 µg of polydI/dC). DNA-protein complexes were separated by electrophoresis. Specific binding was competed with excess of unlabeled oligonucleotides. For antibody super shifts, anti-Sp1, anti-Sp3, anti-AP4, (Santa Cruz Biotechnology, Santa Cruz, CA, USA), and anti-HA (Roche) were incubated with the mixture for 20 min on ice before labeled probes were added. Dried gels were analyzed on a PhosphorImager.
|
Chromatin immunoprecipitation assay
The Chromatin Immunoprecipitation Assay (ChIP) assay kit (Upstate Biotechnologies, Charlotteville, VA, USA) was used with minor modifications. For GR ChIP, cells kept in serum-free medium for 24 h were transfected with rat GR expression plasmid using FuGene6. After 24 h, washed cells were treated with either vehicle or 0.1 µM dexamethasone for 3 h in serum-free medium. For NF1 ChIP, cells were transfected with NF1-A1.1 expression plasmid in normal medium. After 24 h, cells were lysed and DNA fragmented, and 10 µl of the chromatin solution were saved as input. Five micrograms of different antibodies were added to tubes containing chromatin solution. Antibody complexes captured using protein A-agarose beads were pelleted and washed. The chromatin was extracted, reverse cross-linked, DNA purified (DNeasy kit, Qiagen), and amplified using forward 5'-GGTGAGCACCGGCTGGTTCCTCG-3' and reverse 5'-TTTCCTTCACTTCTCTCCCGGCA-3' primers. Bands were visualized using ethidium bromide.
Bioinformatics and statistics
Transcription factor binding sites were analyzed with Match (http://www.generegulation.com/pub/programs.html#match). Values are reported as mean and SD. Unpaired t test, Bonferoni test, and Pearson 
2 test were performed for statistical analyses of the two collectives. The allele distribution was estimated including all allele lengths studied, and the median basepair size was determined. The dichotomization as "short" or "long" allele was performed by using the median of all alleles (13)
. "Short" comprised less than 364 bp (<364) and "long" equal or longer than 364 bp (
364).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
B). Sequencing confirmed four polymorphic sites at positions –209 (G or A), –194 (G or C), –151 (G or A), and –126 (G or A) (Fig. 1A
). Analyses indicated that C instead of G at –194 eliminates an XhoI restriction site. Similarly, A instead of G at –151 results in loss of a HinfI restriction site (Fig. 1B
). After sequencing and analyses for potential transcription factor binding sites, we focused on polymorphisms located at –209 and –126, each revealing a G to A transition (Fig. 1A
). A instead of G (–209 and –126) eliminates an NF1 or a Sp1 binding site, respectively. The G-194C and G-151A polymorphisms did not associate with defined transcription factor binding sites.
|
The frequency of polymorphisms was analyzed in the same 54 salt-sensitive and 96 salt-resistant subjects (Table 2
). The G-209A variant was relatively more often present in the salt-sensitive (5 out of 54, 9.3%) than in salt-resistant subjects (2 out of 96, 2.1%; P<0.05), whereas the G126A variant appeared in a similar percentage of salt-sensitive (9 out of 54, 16.7%) and salt-resistant subjects (11 out of 96, 11.5%). The presence of both variants was observed in one salt-sensitive subject.
|
Genotyping
A total of 15 different alleles of the polymorphic microsatellite marker in intron 1 and 2 was detected. The length of the PCR products varied from 356 nucleotides for allele A1 to 384 for allele A15. Heterozygosity reached 59% for SS and 68% for SR (Table 2)
. SS presented more often A7/A7 variants than SR (24.1 and 9.4%; P<0.01). Interestingly, the NF1 polymorphism G-209A was associated with the A4/A7 variant and was present in four of five SS and in one of two of SR subjects. This A4/A7 variant was observed in only one single subject out of 143 without polymorphism.
The (THF+alpha-THF)/THE ratio was higher in SS than SR (1.31±0.43 vs. 0.95±0.31; P<0.0001; Table 2
), a difference found for both homologous and heterologous alleles (Table 2)
. Allele length was dichotomized as "short" (alleles<364 bp, which is the median length of all alleles) or "long" (364 bp) in the same way as has been previously described (13)
. The group of individuals with short alleles (<364 bp: n=57) had lower (THF+5alpha-THF)/THE ratios than that with long alleles (364 bp: n=93; 0.94±0.35 vs. 1.17±0.40; P<0.005). Subjects with NF1 polymorphism (n=7) had higher (THF+alpha-THF)/THE ratios than subjects with SP1 polymorphism (n=20; 1.43±0.34 vs. 0.97±0.33; P<0.04) but not significantly higher than subjects without mutations (n=123; 1.43±0.34 vs. 1.08±0.4; nonsignificant).
Polymorphisms in the HSD11B2 promoter decrease promoter activity
To understand the role of these polymorphisms with respect to transcriptional regulation, we performed transient transfections of reporter constructs in SW620 cells. The following constructs were used: pGL3–400 (wild type), Mut-(–209) containing A instead of G at position –209, Mut-(–126) containing A instead of G at position –126, Mut-(–194) containing C instead of G at position –194, and Mut-(–151) containing A instead of G at position –151. Promoter activity was analyzed 30–36 h after transfection and normalized to Renilla luciferase. The transfection experiments showed a better expression of wild-type construct than Mut-(–209), Mut-(–126), and Mut-(–151) (Fig. 2
A). Mutation at position –194 [Mut-(–194)] did not affect the expression (Fig. 2A
). Transcriptional search analysis revealed several NF1 and glucocorticoid receptor binding sites (GRE; Fig. 1A
). GR and NF1 interact synergistically to activate the transcription of mouse mammary tumor virus (MMTV; ref 30
). To examine whether this cooperation takes place and whether the common polymorphism at position –209 (G-209A) in the NF1 binding site has any effect on this cooperation, we cotransfected GR-expressing vector with pGL3–400 and Mut-(–209) in presence or absence of different isoforms of NF1 expression plasmids (Fig. 2B
). NF1 alone enhanced promoter activity of wild type and mutant but the activity of wild-type to mutated construct [% of WT in Fig. 2B
] was slightly higher when the four NF1 isoforms were cotransfected. Similarly cotransfection of a GR-expressing vector with wild-type and mutated constructs increased total activity of the promoter in both constructs. Again the percentage ratio of wild-type to mutated construct increased (Fig. 2B
). However, cotransfection of different NF1 isoforms and GR-expressing plasmid with wild-type (pGL3–400) or mutated construct [Mut-(–209)] enhanced promoter activity (Fig. 2B
). Surprisingly the ratio of wild-type to mutated construct activity increased. These results indicate that different NF1 isoforms cooperate with GR to activate HSD11B2 gene promoter, confirming an observation made by Chaudhry et al. (30)
while studying a different gene. A mutation at the NF1 binding site (position –209) not only decreased basal activity (Fig. 2A
) but also destabilized cooperation between GR and NF1. Thus, less promoter activity can be expected from mutated alleles (Fig. 2B
).
|
Polymorphisms in the HSD11B2 promoter are targets for transcription factor binding
Analyses with the TRANSFAC (Transcription Factor Database) program indicated potential binding sites for NF1, Sp1, and GRE overlapping the two polymorphisms G-209A and G-126A (Fig. 1A
). Transition of G to A (–209 and –126) eliminates a NF1 and Sp1 binding site. To study the effect of these mutations on protein-DNA interactions, we performed electrophoretic mobility shift assay (EMSA) with wild-type and mutated DNA probes (–227 to –200) using SW620 cell nuclear protein extracts after overexpressing hemagglutinin tagged NF1-X2. The probes contained either G in W-(–209) or A in Mut-(209) nucleotide at position –209. W-(–209) has a potential binding site for both Sp1 and NF1, whereas Mut-(209) only for Sp1. Two specific complexes were formed with wild-type probe [W-(–209)] and one was formed with mutated probe [Mut-(–209)] (Fig. 3
A). Total displacement of both complexes formed was achieved with increasing amounts of wild-type probe [W-(–209)-5x, W-(–209)-10x, W-(–209)-25x, and W-(–209)-50x]. Mutated competitor competed only with one of the complexes indicating that polymorphism at NF1 binding site (G-209A) decreases its affinity for NF1 (Fig. 3A
). Incubating nuclear proteins with antibody against hemagglutinin-tag (Ab.HA) inhibited the lower complex formation, while antibody against Sp1 (Ab.Sp1) inhibited the upper complex formation, indicating that lower complex is NF1 and upper complex is Sp1 (Fig. 3A
). Similarly, the polymorphism at position –126 (G-126A) was analyzed using two sets of wild-type and mutated probes (region –144 to –118; Fig. 3B
). This binding of Sp1 and Sp3 transcription factors to their recognition-binding site was significantly impaired when transition of G to A occurred at position –126 (Fig. 3B
).
|
GR directly interacts with the HSD11B2 promoter
To determine whether GR directly interacts with the HSD11B2 promoter, we performed EMSAs using two sets of probes containing GR binding sites GRE1 (–372 to –343) or GRE2 (–176 to –163; Fig. 1A
) and nuclear extracts from MCF-7 after dexamethasone treatment (0.1 µM, 6 h). Two shifted bands in both GRE1 and GRE2 radiolabeled probes were observed (Fig. 3C
). The specificity of binding was confirmed by displacement experiments (Fig. 3C
). Incubating NE with anti-GR antibody (Fig. 3C
; Ab.GR) but not unrelated AP4 antibody (Fig. 3C
; Ab.AP4) blocked complex formation confirming that the complex is GR. These results suggest that GR forms complexes directly with HSD11B2 promoter.
GR needs NF1 to cooperatively activate the HSD11B2 promoter
To determine which regions of the HSD11B2 proximal promoter are necessary for glucocorticoid-mediated transcriptional activation, we used pGL3–400 comprising all putative GREs (–375 to +122) and pGL3-Xho (–197 to +122) luciferase reporter vectors controlled by the HSD11B2 promoter (Fig. 1A
). In reporter constructs pGL3–400(GRE-) and pGL3-Xho(GRE-), the most proximal GRE binding site (GRE2; –184 to –162) was disrupted and pGL3-Xho(GRE-) had no GRE (Fig. 1A
).
To test the effect of GR and/or NF1 binding on the HSD11B2 promoter activity, these four constructs were cotransfected with GR and/or NF1-A1.1 expression vectors in either SW620 or LLCPK1 cells (Fig. 4
A, B). Luciferase activity was determined in transiently transfected cells after 24 h incubation with vehicle or 0.1 µM of dexamethasone. Figure 4A
depicts that NF1 alone enhances the promoter activity of all wild-type but less of GRE2 mutated constructs.
|
Cotransfection of a GR expressing vector with wild-type and GRE2 mutated constructs in SW620 cells increased the total activity of the promoter using constructs containing GRE2 binding site(s) (pGL3–400 and pGL3-Xho) more than vectors with mutated GRE2 [pGL3–400(GRE-); pGl3-Xho(GRE-); Fig. 4A
]. Interestingly, cotransfection of the NF1-A1.1 isoform and GR-expressing plasmids with wild-type or GRE2 mutated constructs enhanced promoter activity synergistically. Similarly in LLCPK1 cotransfections of a GR-expressing vector with wild-type or GRE2 mutated constructs increased the total activity of the promoter in those constructs with at least one GRE binding site(s) [pGL3–400, pGL3-Xho, and pGL3–400(GRE-); Fig. 4B
]. These results provide functional evidence for the presence of GREs in the proximal region of the HSD11B2 promoter. Comparing the results from SW620 and LLCPK1 cells after cotransfection of a GR-expressing vector with wild-type or GRE2 mutated constructs (Fig. 4A, B
) suggests cell-type specific importance of GRE1 and GRE2 and, furthermore, that the region between –375 to –197 is necessary for cooperation between NF1 and GR. Since the region (–375 to –197) overlapped with the polymorphism (–209, Fig. 1A
), the observations depicted in Fig. 4
are in line with those in Fig. 2
, indicating that NF1 cooperates with GR to activate HSD11B2 gene promoter.
Dexamethasone treatment increases HSD11B2 expression and activity
To investigate the potential role of glucocorticoids regulating HSD11B2 expression, MCF-7 and Ishikawa cells were treated with vehicle, 0.1 µM dexamethasone, or 1 µM RU38486 (RU486) for up to 48 h in serum-free medium. Dex enhanced HSD11B2 gene expression and activity time dependently (Fig. 5
A, B). RU38486 completely blocked this induction (Fig. 5C, D
), a finding in agreement with luciferase induction (Figs. 2
and 4)
observed using Ishikawa and BEAS-2B cells (32
, 33)
.
|
Endogenous HSD11B2 expression and activity increase after over-expression of GR and NF1
The cooperation of GR and NF1 regulating endogenous HSD11B2 expression was studied by over-expressing NF1 (NF1-A1.1, NF1-X2) without or with GR-expressing vector (Fig. 6
). After transfection, SW620 cells were grown for additional 48 h with 0.1 µM dexamethasone. NF1 in combination with GR increased mRNA and activity of HSD11B2 (Fig. 6)
. This confirms a synergistic effect of NF1 and GR considering endogenous HSD11B2.
|
GR, NF1, and Sp1 interact with HSD11B2
Computer analysis of the proximal region of the HSD11B2 5'-regulatory region revealed numerous high score NF1, GRE, and Sp1 binding site(s) (Fig. 1A
). Using EMSA, we showed (Fig. 3)
that these NF1, GRE, and Sp1 binding sites are recruited by their cognate transcription factors. To demonstrate that these transcription factors form a complex with corresponding binding sites in vivo, a ChIP assay was performed using cross-linked chromatin from MCF-7 and SW620 cells. Figure 7
demonstrates that NF1, Sp1, and GR complexed with the DNA region of HSD11B2 promoter after dexamethasone treatment.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. The HSD11B2 promoter contains relevant NF1 half-sites for NF1-induced activation as shown recently by our group (17)
. The G-209A polymorphism is located within such a half-site. The corresponding construct exhibits a decreased basal and NF1-induced activity, an observation paralleled by a reduced binding affinity of NF1. Additionally, the G-209A polymorphism has a decreased affinity for Sp1 paralleled by a decreased expression, a finding in line with the concept that Sp1 is relevant for basal expression of HSD11B2 (21
, 22)
.
Dexamethasone treatment up-regulates HSD11B2 expression in native cells (Fig. 5)
, an effect that is deployed by GR, because a GR antagonist reversed the dexamethasone action (32
, 35)
. An enhancing effect on the GR-mediated dexamethasone effect by NF1 was identified by overexpressing GR and/or NF1 (Fig. 6)
. Such cooperative behavior between NF1 and GR has previously been proposed for other genes (30
, 36
, 37)
. The half-site consensus sequence (TGTTNT) can bind to GR (38)
. Such GRE half-sites are present on the proximal region of the HSD11B2 promoter and recruit GR as shown by EMSA in vitro and by ChIP assay in intact cells (Fig. 3C
and Fig. 7
). Interestingly, the NF1 binding site comprising the G-209A polymorphism is located in the proximity of several GREs (Fig. 1A
). A synergism between NF1 and GR can be identified (Fig. 4)
, an observation in agreement with a recent study showing that a mutation in the NF1 binding site reduces both basal and dexamethasone-dependent transcription of the MMTV promoter (39)
. The latter investigations suggested furthermore that NF1 may participate in chromatin remodeling in addition to directly enhancing transcription and that in the absence of its binding site, the GR is unable to effectively bind the promoter (39)
. Thus, one might predict that the effect of a mutation in a NF1 binding site at a location where it interacts with a GRE as shown by our transient transfection studies might be even more pronounced in humans harboring the G-209A polymorphism where the chromatin structure exists.
Sp1 and Sp3 have previously been shown to be relevant for basal HSD11B2 expression (18
, 21
, 22
, 40)
. The G-126A polymorphism, which results in a decreased transcriptional activity (Fig. 2A
), is within an Sp1 binding site as demonstrated by EMSA and ChIP assays (Fig. 3B
and Fig. 7
). The G-126A variant is in proximity with NF-B sites (21
, 22)
. NF-B can act synergistically with Sp1 on certain promoters (41)
and has been shown by us to modulate 11beta-HSD2 expression (22)
.
Associations between a polymorphic microsatellite marker of the HSD11B2 gene and the 11beta-HSD2 activity and/or blood pressure, salt sensitivity, or response to thiazide diuretics have been reported previously (10
, 12
, 13
, 24
, 25
, 42)
. In the present investigation, CA repeat length was associated with (THF+5alpha-THF)/THE ratio, a finding in line, first, with the observation that CA repeat length was correlated with the (THF+5alpha-THF)/THE ratio in 377 genetically homogenous essential hypertensives from North Sardinia (24)
and, second, with the report of a cohort of 33 subjects with low renin hypertension showing that longer CA alleles were found in patients with relatively low 11beta-HSD2 activity, and third, with results derived from cell culture experiments demonstrating that transfection of minigenes containing 23 instead of 14 CA repeats diminished the expression by 50% (13)
. Correlations between microsatellite markers and enzyme activity in the absence of mutations in the coding region suggest, among other mechanisms, the presence of variants in the promoter. Here, we identified such variants and present evidence for variable promoter activity.
Recently, many reports described genetic polymorphisms in patients with salt sensitivity or hypertension. However, virtually no mechanistic studies have addressed the molecular mechanisms controlling the expression of such genes by deciphering the interaction between polymorphisms in the promoter and transcription factors (43)
. The present investigation demonstrates that such interactions, including the modulation of cooperation between nuclear factors as exemplified here for NF1 and GR, are important for transcription, as shown by the higher frequency of the G-209A polymorphism in salt-sensitive than salt-resistant subjects. Although these findings appear statistically significant, their quantitative importance for explaining salt sensitivity and/or hypertension is limited, as this has generally been the case in the area of human molecular genetics of hypertension (44)
. Nevertheless, our findings agree with the concept that the identification of subtle genetic changes with defined functional relevance might be important for restoring the credibility of the field of genetics of hypertension, which has been undermined by the huge number of association studies with contradictory results (42)
.
In conclusion, polymorphisms were identified for the first time in the promoter of the HSD11B2 gene. These variants reduce significantly the activity of the cognate transcription factors, resulting in diminished transcription of HSD11B2 with a possible association with salt sensitivity.
|
ACKNOWLEDGMENTS |
|---|
Received for publication March 19, 2007. Accepted for publication May 3, 2007.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
