The CAG Repeat Polymorphism in the AR Gene

0013-7227/01/$03.00/0 Printed in U.S.A.

The Journal of Clinical Endocrinology & Metabolism 86(10):4867– 4873 Copyright © 2001 by The Endocrine Society

The CAG Repeat Polymorphism in the AR Gene Affects High Density Lipoprotein Cholesterol and Arterial Vasoreactivity ¨ RG GROMOLL, MICHAEL ZITZMANN, MAIK BRUNE, BRITTA KORNMANN, JO SIGRID VON ECKARDSTEIN, ARNOLD VON ECKARDSTEIN, AND EBERHARD NIESCHLAG Institutes of Reproductive Medicine (M.Z., M.B., B.K., J.G., S.v.E., E.N.), Clinical Chemistry and Laboratory Medicine (A.v.E.), and Arteriosclerosis Research (A.v.E.), University of Mu¨nster, D-48129 Mu¨nster, Germany Genomic effects of T are exerted via the AR. The length of the polymorphic CAG repeat sequence in the AR gene is inversely correlated with the transcriptional regulation of target genes by T. In 110 healthy men (20 –50 yr), we investigated the interactions among this polymorphism, serum levels of sex hormones, cardiovascular risk factors, and flow-mediated and nitrate-induced vasodilatation of the brachial artery. The number of CAG repeat had no significant correlations with serum concentrations of total or free T. Stepwise multiple regression analysis revealed positive correlations of the number of CAG repeat with serum levels of high density lipoprotein cholesterol (partial r ⴝ 0.44; P < 0.001) and flow-mediated vasodilatation (partial r ⴝ 0.37; P < 0.001). The association of

L

ITTLE IS KNOWN about the role of T in atherosclerosis, which was implicated of contributing to the approximately 10-yr gap in the clinical manifestation of coronary heart disease (CHD) between men and women. In epidemiological and clinical studies, plasma levels of T had significant associations with the incidence or presence of CHD, which, however, were inverse in men and positive in women (1–3). Moreover, hypoandrogenemia in men and hyperandrogenemia in women are confounded with various metabolic disorders, including obesity, insulin resistance, dyslipidemia, and impaired fibrinolysis (3). Finally, case-control studies are limited by the problem that chronic illnesses, including CHD, as a stressful condition modulate serum levels of T (4). Few data are available from interventional studies: abuse of anabolic androgens has been held responsible for premature myocardial infarction in some athletes (5, 6); the contrary was observed when application of T to male CHD patients reduced clinical, electrocardiographic, and angiographic signs of coronary ischemia (1, 3, 7–9). In fact, T exerts both beneficial and adverse effects on cardiovascular risk factors by decreasing serum levels of lipoprotein(a), fibrinogen, and high density lipoprotein (HDL) cholesterol (1, 3). Likewise, conflicting data have been reported on the effect of T on vascular functions. Androgen deprivation in men was found to improve endothelial function (10, 11), and Abbreviations: apoA-I, Apolipoprotein A-I; CHD, coronary heart disease; CV, coefficient of variation; FFM, fat-free mass; FMD, flow-mediated vasodilatation; GTN, glycerol trinitrate; HDL, high density lipoprotein; LDL, low density lipoprotein; sE-selectin, soluble E-selectin; sVCAM-1, soluble vascular cell adhesion molecule 1; TBW, total body water.

CAG repeat with high density lipoprotein (HDL) cholesterol was independent of body fat content and serum levels of free T, which both had significant negative correlations with HDL cholesterol. The association of CAG repeat with flow-mediated vasodilatation was independent of cigarette smoking and serum levels of free T and low density lipoprotein cholesterol, which also were correlated with flow-mediated vasodilatation. We conclude that a low number of CAG repeat in the AR gene implies a greater chance for low levels of HDL cholesterol and reduced endothelial response to ischemia, which are both important risk factors for coronary heart disease. (J Clin Endocrinol Metab 86: 4867– 4873, 2001)

high dose application of androgens had detrimental effects in transsexual women (12); discrepantly, acute intravascular administration of T at very high dosages improved coronary artery relaxation and flow-mediated vasodilatation (FMD) of the brachial artery in men (13); low dose T treatment in postmenopausal women also ameliorated endothelial functions (14). Animal studies suggest that T regulates vasoreactivity by endothelium-dependent and -independent mechanisms, but are contradictory as well (15–20). The genomic effects of T are exerted via activation of the AR, which regulates the transcription of target genes. A variable number of CAG repeats in exon 1, which encodes for glutamine residues in the amino-terminal domain of the AR and which normally ranges between 9 –35 (21, 22), is inversely associated with the transcriptional response to T (23, 24). As the clinical consequence, the number of CAG repeats was shown to be associated with the incidence of prostate cancer (21, 22, 25, 26), benign prostatic hyperplasia (27, 28), and impaired spermatogenesis (29 –32). Furthermore, abnormal expansion of the CAG repeat length leads to Kennedy’s disease, which is accompanied by morphological hypoandrogenic traits (33–35). In this study of healthy males, aged 20 –50 yr, we addressed the question of whether the CAG repeat polymorphism in the AR gene influences cardiovascular risk factors and endothelial functions that may be impaired before the clinical onset of CHD (36 – 40). Subjects and Methods Besides the assessment of lipid profiles, life-style factors, body composition, and sexual hormones, we determined the CAG repeat poly-

4867

4868

J Clin Endocrinol Metab, October 2001, 86(10):4867– 4873

morphism in the AR and investigated endothelial functions as putative early marker of vascular disease by noninvasively measuring the FMD of the brachial artery, a method previously demonstrated to be associated with coronary endothelial function and cardiovascular risk factors (38, 41– 47).

Subjects Healthy male Caucasians, aged 20 –50 yr, were recruited by local newspaper advertisements that asked for volunteers willing to participate in a 1-term clinical study including the assessment of lipids, body fat content, and sexual hormones. They were informed that a genetic polymorphism was also to be investigated. They received a payment of 20 German marks and were all interested in obtaining the results. The study was approved by the ethics committee of the University and the State Medical Board (Munster, Germany; no. 0VIINie1). All volunteers gave written informed consent. After exclusion of previous androgen use, atherosclerosis, diabetes mellitus, arterial hypertension, dyslipoproteinemia, medication of any kind, drug abuse, a possible state of androgen deficiency (total serum T ⬍12 nmol/liter or free T ⬍250 pmol/liter), alcohol intake of more than 40 g/d, and renal or hepatic illness by history, physical examination, and serum/blood analysis, 110 men were eligible to participate in the study. Eight men were excluded for taking medication for arterial hypertension, 3 were excluded for use of cholesterol-lowering drugs, 2 were excluded for extensive consumption of alcohol, and 1 was excluded for previously diagnosed and treated diabetes mellitus type 1. No exclusion was necessary because of abnormal laboratory findings or a physical exam revealing pathological entities of any kind. Habitual cigarette smoking, alcohol consumption (grams per d), body weight, body height, and body mass index were assessed. Physical activity was expressed as the amount of strength and/or endurance training in hours per wk, resulting in an overall score of 1 (none) to 6 (⬎9 h/wk). This information was obtained by a standardized interview [retesting interobserver coefficient of variation (CV) in 18 subjects after 3 months was 8.5%]. All examinations and blood sampling were performed between 0800 –1000 h after an overnight fast, including abstinence from caffeine-based drinks for 12 h. The clinical part of the study was completed within 4 wk during the summer.

Vascular ultrasound Doppler ultrasound was used to assess the vascular reactivity of the brachial artery. Endothelium-dependent vasodilatation was induced by hyperemia, which triggers the release of nitric oxide from endothelial cells (48). For the assessment of endothelium-independent vasodilatation, the endothelium was bypassed by the exogenous application of glycerol trinitrate (GTN) (49). The method of assessing endothelial functions noninvasively was previously validated in studies of atherosclerosis (38, 43, 50) and is suitable for detecting the influence of androgens on the vascular endothelium (10 –14). Before the start of the examination the subjects rested in a room with a temperature of 20⫺24 C for 10 –15 min. Subjects were investigated in the supine position. A high resolution ultrasonography 7.5-MHz linear phased array ultrasound transducer (ultrasound scanner type 2002 ADI, B-K Medical, Gentofte, Denmark) was used to image the dominant arm brachial artery longitudinally just above the antecubital fossa. The artery was identified by a pulsed Doppler signal at a 70° angle to the vessel with the range gate (1.5 mm) in the center of the artery. Additionally, colorcoded duplex sonography was used. B-mode ultrasound images were used for measurements, the transmit zone was set to the depth of the near wall, and gain settings were not changed during the study. Digitalized motion sequences were saved for analysis. Arterial diameter was assessed over a 1-cm straight segment by measuring the distance from the anterior to the posterior wall (m-line, the interface between media and adventitia) at maximal resolution using the implemented software for distance measurements. After baseline images were obtained, a blood pressure cuff was placed over the ipsilateral upper arm just above the transducer, inflated for 4 min at 180 mm Hg, then suddenly deflated. Images of the flow-mediated dilator response were obtained 1 min after cuff release (maximal arterial dilatation). The brachial artery diameter was then allowed to return to normal (8 min). New baseline images were recorded, then 0.4 mg sublingual GTN was administered. Images from

CAG Repeat in AR Gene and HDL Cholesterol • Zitzmann et al.

endothelial-independent vasodilation were obtained 3 min after application. The vessel diameter was assessed during maximal and minimal blood flow, as indicated by the color-coded pulse wave of three cardiac cycles at each time point; respective values were then averaged. This method captures the pulse-dependent maximum and minimum changes in vessel diameter. Body size (and, hence, distance from the aortic valve to the location of diameter assessment) does not influence the results as might be the case in electrocardiogram-triggered measurements. Flowmediated and GTN-induced vasodilatation were calculated as the percent change in diameter compared with baseline. All scans were performed by one investigator who was blinded with respect to the subjects’ smoking habits, DNA analysis, or levels of lipids and sexual hormones. Respective intraobserver variability for repeated measurements of the vessel diameter is 1.6%. When reactive hyperemia studies were performed on 2 separate d, the intraobserver CV for FMD was 2.6% (based on 19 subjects).

Sexual hormones Serum T levels were measured with a commercial ELISA kit (DRG Instruments GmbH, Marburg, Germany); serum levels of LH, FSH, PRL, prostate-specific antigen, SHBG, and E2 were measured with highly specific time-resolved fluoroimmunoassays (Autodelfia, Freiburg, Germany). The mean intraassay CV was below 5%; the mean interassay CV was below 10%. Levels of free T were calculated from levels of SHBG and total serum T according to the law of mass action, using 3.6 ⫻ 104 liter/mol as the association constant of T with albumin and 1 ⫻ 109 liter/mol with SHBG. Calculation with this method yields highly reliable values of levels of free T (51).

Biochemical analyses A Hitachi 917 autoanalyzer was used for the quantification of serum concentrations of triglycerides and cholesterol with enzymatic tests, of HDL cholesterol with a homogenous assay, and of apolipoprotein A-I (apoA-I) with and lipoprotein(a) with (latex-enhanced) turbidimetric immunoassays (Hitachi/Roche, Mannheim, Germany). Imprecision was below 5%. Low density lipoprotein (LDL) cholesterol was calculated using the Friedewald formula (52). Plasma concentrations of soluble E-selectin (sE-selectin) and soluble vascular cell adhesion molecule 1 (sVCAM-1) were determined in duplicate by the use of enzyme immunoassays from Bender Medsystems Diagnostics GmbH (Vienna, Austria). The interassay CV was below 10%.

Determination of the number of CAG repeats within exon 1 of the AR gene DNA was isolated from EDTA blood samples using the Nucleon Kit (Amersham Pharmacia Biotech, Freiburg, Germany). A fragment of exon 1 of the AR was amplified by PCR. Each reaction sample (25 ␮l) contained 100 –500 ng genomic DNA, 20 pmol AR exon 1 for primer (5⬘-GCCTGTTGAACTCTTCTGAGC-3⬘), 20 pmol AR exon 1-rev primer [5⬘-CGATGGGCTTGGGGAGAACCATCCTCA-3⬘, IRD (infrared dye)800 labeled], reaction buffer [10 nmol/liter Tris-HCl (pH 8.3), 50 nmol/ liter KCl, 0.01% gelatin, 2 nmol/liter MgCl2, and 0.2 mmol/liter deoxyNTPs], and 2 U Taq polymerase (Promega Corp., Heidelberg, Germany). Denaturation at 94 C for 50 sec was followed by annealing at 58 C for 40 sec and an extension step at 72 C for 1 min. After initial incubation at 94 C for 2 min, PCR was performed for 35 cycles, with a final extension step at 72 C for 10 min. PCR products were tested by 2% agarose gel electrophoresis. The fluorescent PCR products (1–2 ␮l) were diluted with 2 ␮l loading dye fluorescent buffer sample (Amersham Pharmacia Biotech), and the total volume was adjusted to 5 ␮l using aqua dest. The samples were heated to 70 C for 1 min, then chilled on ice, and 1–2 ␮l were loaded onto a 6% denaturing sequencing gel. Samples were electrophoresed (Licor 4200, LiCor, Inc., Lincoln, NE) at 1500 V for 8 h. The number of CAG repeats were calculated by comparing the detected PCR fragment to sequencing reactions, which were run in parallel to the samples and served as molecular size markers. For example, a fragment with 372 bp contains 21 CAG repeats. In addition, PCR products with known numbers of CAG repeats (15, 21, and 30 CAG repeats, as determined by cloning and sequencing of corresponding AR’s exon 1 fragments) were assessed during each determination as well and were


J Clin Endocrinol Metab, October 2001, 86(10):4867– 4873 4869

used as internal standards for the calculation. By this, migration aberrations due to changes in the electrophoretic properties, which could result in incorrect repeat calculations, were excluded. One gel showing irregular bands was discarded. Determination of CAG number was repeated twice on two separate gel runs. Discrepancies were not observed. Eighteen random samples were cloned and sequenced to confirm the validity of the measurement method being used.

Bioimpedance measurements Body fat content was assessed by bioimpedance measurement using a B.I.A 2000-M multifrequency body composition analyzer (Data Input, Inc., Frankfurt/Main, Germany). A multifrequent alternating current (1, 5, 50, or 100 kHz, respectively) was used to eliminate variations in extraand intracellular water contents. This method is more accurate than a single frequency approach (53). The same apparatus was used throughout, following the instructions given by the manufacturer. The patients were examined in a supine position after voiding. Total body water (TBW) was calculated from the measured impedance using the following formula: TBW ⫽ height (2)/impedance ⫻ 0.585 ⫹ 1.825 (54). Fat-free mass (FFM) was then calculated from the TBW by assuming 73.2% hydration of FFM: FFM ⫽ TBW/0.732 (55). Body fat mass was calculated using a two-compartment model: body weight ⫽ FFM ⫹ BFM. In our institute, the intraobserver CV is 2.34%; the interobserver CV is 6.77% (based on measurements by 3 investigators in 15 subjects). The measurements were performed by one investigator.

Statistics All variables were checked for normal distribution by the Kolmogorov-Smirnov one-sample test for goodness of fit. When necessary, analysis was performed logarithmically or, for percentage values, on arcsine-transformed data. Basic correlations of the number of CAG repeats with parameters of interest were calculated after adjustment of data for age and body fat content. To investigate the simple influence of cigarette smoking on FMD, a two-tailed t test for unpaired samples was applied. To study the influence of T and the AR polymorphism on lipid profiles, FMD, and GTN-induced vasodilatation, the respective parameter was introduced as the dependent variable into a stepwise multiple regression model. The number of CAG repeats; levels of free T and E2, and other parameters of possible influence were entered as controlling variables. For lipid profiles, the controlling variables were body fat content, log age, physical activity, log alcohol consumption, and cigarette smoking; these controlling variables were applied for vascular parameters as well. Additionally, levels of LDL cholesterol and the baseline diameter of the brachial artery (showing a significant negative correlation with FMD in previous studies (38, 47) were entered. Computations were performed using the statistical software package SPSS (release 9.0.1, SPSS, Inc., Chicago, IL). Unless otherwise stated, results are given as the mean ⫾ sd in tables and figures. A two-sided P ⬍ 0.05 was considered significant. Levels of statistical significance are shown in the figures as asterisks (*, P ⬍ 0.05; **, P ⬍ 0.01; ***, P ⬍ 0.001).

Results General data

The range and distribution of the number of CAG repeats are displayed in Fig. 1; data for other variables are given in Table 1. The number of CAG repeats was significantly positively correlated with levels of LH, as described previously (56) (Table 2). The AR polymorphism had no significant association with serum levels of free T (Table 2). The number of CAG repeats showed a significant positive basic correlation with FMD, GTN-induced vasodilatation, and HDL cholesterol levels as well as a slight negative correlation with serum levels of triglycerides (Table 2). Of 110 participating men, 66 were nonsmokers, and 44 were habitual smokers (⬎5 cigarettes/d). FMD was lower in smokers than in nonsmokers (13.54 ⫾ 6.27% and 9.26 ⫾ 5.89%; P ⬍ 0.001, by two-tailed unpaired t test).

FIG. 1. The distribution of the number of CAG repeats within exon 1 of the AR gene in 110 healthy males. TABLE 1. General data Parameter

Anthropometrical data Age (yr) Body mass index (kg/m⫺2) Body fat content (%) Body ht (cm) Body wt (kg) Sociometrical data Physical activity score Nonsmokers Smokers Alcohol consumption (g/week) Hormones Total T (nmol/liter) Free T (pmol/liter) E2 (pmol/liter) LH (U/liter) CAG repeats in the AR gene (exon 1) Vascular functions Baseline diameter of the brachial artery (mm) Flow-mediated dilatation (%) Glycerol trinitrate-induced dilatation (%) sVCAM-1 (ng/ml) sE-Selectin (␮g/ml) Lipid profiles Total cholesterol (mmol/liter) LDL cholesterol (mmol/liter) HDL cholesterol (mmol/liter) Triglycerides (mmol/liter) Lp(a) (g/liter) ApoA-1 (mg/dl)

Value

20 –50, 28 23.8 ⫾ 3.0 20.9 ⫾ 5.3 182.3 ⫾ 7.1 79.3 ⫾ 11.6 2.5 ⫾ 0.9 66 44 0 –280, 60 21.4 465.2 67.1 3.9 21.62

⫾ ⫾ ⫾ ⫾ ⫾

5.7 126.7 18.0 1.8 3.14

4.2 ⫾ 0.5 11.6 ⫾ 6.2 17.6 ⫾ 7.0 10.64 ⫾ 2.65 1.53–19.31, 5.96 4.79 ⫾ 0.99 2.76 ⫾ 0.87 1.49 ⫾ 0.33 1.23 ⫾ 0.87 0.03–1.49, 0.12 131.8 ⫾ 16.5

Data are given as the mean ⫾ SD in case of normal distribution, otherwise as the range and median.

Lipid profiles

By stepwise multiple regression analysis, the number of CAG repeats showed a positive and highly significant correlation with serum levels of HDL cholesterol (Table 3), which is independent of body fat content and levels of free T (both negative association; Table 3). The AR polymorphism accounted for 14.0% of the total variance in HDL cholesterol (partial r ⫽ 0.44; P ⬍ 0.001; Fig. 2). Figure 3 displays levels of HDL cholesterol according to the quartiles of the CAG repeat distribution. The HDL cholesterol levels of men within the lowest quartile of CAG repeat numbers were significantly lower (mean ⫾ sd, 1.722 ⫾ 0.29 mmol/liter) than the HDL cholesterol levels of men within the highest

4870



TABLE 2. Basic correlations Parameter

Basic correlations with the no. of CAG repeats in the AR gene

Hormones Total T Free T E2 FSH LH Vascular functions Baseline diameter of the brachial artery Flow-mediated vasodilatation Glycerol trinitrate-induced vasodilatation sVCAM-1 sE-Selectin Serum lipids Total cholesterol LDL cholesterol HDL cholesterol Triglycerides Lp(a)

r ⫽ 0.138 r ⫽ 0.166 r ⫽ 0.023 r ⫽ 0.095 r ⫽ 0.226; P ⫽ 0.019

NS NS NS NS

r ⫽ ⫺0.149 r ⫽ 0.324; P ⫽ 0.001 r ⫽ 0.220; P ⫽ 0.021 r ⫽ 0.012 r ⫽ ⫺0.038

NS NS NS

r ⫽ 0.007 r ⫽ ⫺0.033 r ⫽ 0.241; P ⫽ 0.019 r ⫽ ⫺0.201; Tr, P ⫽ 0.051 r ⫽ 0.094

NS NS NS NS

Parameters were adjusted for age. Correlations: two-sided P ⬍ 0.05 was considered significant. Levels of statistical significance are given for P values below 0.05, 0.01, and 0.001. Tr, Trend (0.05 ⬍ P ⬍ 0.06); NS, nonsignificance (P ⬎ 0.05). TABLE 3. Stepwise multiple regression model for predicting serum levels of HDL cholesterol Variable

Predictor

Standard coefficient (␤)

t

Significance (P)

Overall r2 (%)

HDL cholesterol

CAG repeats Body fat content Free T

0.401 ⫺0.331 ⫺0.240

4.64 ⫺3.87 ⫺2.76

⬍0.001 ⬍0.001 ⬍0.01

34.0 P ⬍ 0.001

Excluded variables: serum E2, cigarette smoking, log age, physical activity, log alcohol consumption, and cigarette smoking. The results of the stepwise multiple regression model concerning levels of apoA-I are given in the text.

FIG. 2. Graphic display of the partial correlation between the number of CAG repeats in exon 1 of the AR gene with serum levels of HDL cholesterol, as derived from the multiple regression model in Table 3 (r ⫽ 0.44; P ⬍ 0.001). HDL cholesterol levels were adjusted for the influence of body fat content and free T according to the multiple regression model and thus appear higher than the raw data (see Table 1).

quartile of CAG repeat numbers (2.024 ⫾ 0.41 mmol/liter; P ⬍ 0.01). In addition, stepwise multiple regression showed the CAG repeat polymorphism to be positively associated with serum levels of apoA-I, which is the main protein component of HDL (␤ ⫽ 0.239; P ⫽ 0.012). This observation was independent of body fat content (␤ ⫽ ⫺0.201; P ⫽ 0.033) and levels of free T (␤ ⫽ ⫺0.175; P ⫽ 0.070). The association between the number of CAG repeats and serum levels of triglycerides just missed the level of statistical significance (Table 4). Levels of LDL cholesterol were best predicted by a multiple regression model including body fat content (␤ ⫽

FIG. 3. Serum levels of HDL cholesterol (mean ⫾ SD, corrected for body fat content) in the quartiles of the CAG repeat distribution (14 –19 CAG repeats, n ⫽ 25; 20 –21 CAG repeats, n ⫽ 35; 22–23 CAG repeats, n ⫽ 21; 24 –31 CAG repeats, n ⫽ 29). Levels are significantly different (by ANOVA, overall P ⫽ 0.0067; by post-hoc test NewmanKeuls test, difference between the highest and lowest quartile is significant at P ⬍ 0.01).

0.323; P ⬍ 0.001), cigarette smoking (␤ ⫽ 0.320; P ⫽ 0.001), serum levels of E2 (␤ ⫽ ⫺0.369; P ⬍ 0.001), and log alcohol consumption (␤ ⫽ ⫺1.99; P ⫽ 0.031; total r2 ⫽ 0.353; total P ⬍ 0.001). Vascular functions

Upon stepwise multiple regression analysis the number of CAG repeats model had a positive and highly significant association with FMD, which is a measure of endothelial function (Table 5). This association was independent of cigarette smoking and serum levels of LDL cholesterol and free



TABLE 4. Stepwise multiple regression model for predicting serum levels of triglycerides Variable

Predictor


t

Significance (P)

Overall r2 (%)

Triglycerides

CAG repeats Body fat content

⫺0.185 0.358

⫺1.96 ⫺3.87

0.053 ⬍0.001

15.5 P ⬍ 0.001

Excluded variables: serum E2, serum free T, cigarette smoking, log age, physical activity, log alcohol consumption, and cigarette smoking.

TABLE 5. Stepwise multiple regression model for predicting flow-mediated dilatation (FMD) of the brachial artery (endotheliumdependent vasodilatation) Variable

FMD

Predictor


t

Significance (P)

Overall r2 (%)

CAG repeats Cigarette smoking LDL cholesterol Free T

0.334 ⫺0.271 ⫺0.456 0.182

3.76 ⫺2.79 ⫺4.41 2.11

⬍0.001 ⬍0.01 ⬍0.001 ⬍0.05

48.0 P ⬍ 0.001

Excluded variables: serum E2, log age, physical activity, log alcohol consumption, and baseline diameter of the brachial artery.

T, which also were correlated with FMD (Table 5). The AR polymorphism accounted for 12.9% of the total variance (partial r ⫽ 0.37; P ⬍ 0.001; Fig. 4). These results were corroborated with high significance when only nonsmoking individuals were introduced into the regression model. Replacement of body fat content by body mass index also confirmed results (data not given). GTN-induced and, hence, endothelium-independent vasodilatation was also significantly and positively correlated with the number of CAG repeats (␤ ⫽ 0.181; P ⫽ 0.049). Negative predictors were cigarette smoking (␤ ⫽ ⫺0.191; P ⫽ 0.036) and the baseline diameter of the brachial artery (␤ ⫽ ⫺0.252; P ⫽ 0.007). Plasma levels of sE-selectin had a significant negative association with FMD (r ⫽ ⫺0.262; P ⫽ 0.006), but not with GTN-induced endothelium-independent vasodilatation (r ⫽ ⫺0.08; P ⫽ NS). These findings are in agreement with its role as a biochemical marker of endothelial function. However, serum levels of neither sE-selectin nor VCAM-1 had any significant association with the number of CAG repeats in the AR gene (r ⫽ ⫺0.038; P ⫽ NS). Discussion

The variable number of CAG repeats in exon 1, which encodes a polyglutamine tract in the amino-terminal domain of the AR, influences transcriptional activity and thereby the response of prostate and testes to T. In this study we extended these observations to nongonadal tissues. The lower the number of CAG repeats, the lower the serum levels of HDL cholesterol and apoA-I and the less extensive was FMD of the brachial artery. This association is also valid with regard to endothelium-independent vasodilatation induced by GTN. Physical activity did not have an influence on these parameters in our investigation. It is debatable whether this could be due to the nature of an ad hoc categorization, although it has a good reproducibility (see Subjects and Methods). As AR variants with a low number of glutamine residues are more responsive to T than AR variants with a long polyglutamine stretch, our findings are in good agreement with the well known HDL-lowering effect of T. This phenomenon and our corresponding results underline that pivotal genes

FIG. 4. Graphic display of the partial correlation between the number of CAG repeats in exon 1 of the AR gene with FMD of the brachial artery (endothelium-dependent vasodilatation) as derived from the multiple regression model in Table 5 (r ⫽ 0.37; P ⬍ 0.001). FMD values were adjusted for the influence of LDL cholesterol, free T, and smoking according to the multiple regression model and thus appear higher than the raw data (see Table 1).

of HDL metabolism are transcriptionally regulated by the AR. Key candidates are hepatic lipase and the scavenger receptor B1, which are both importantly involved in the hepatic removal of HDL lipids (57). Twin studies suggest that 30 –50% of the interindividual variation in HDL cholesterol is explained by genetic variation. However, little is known about the molecular basis of this variation. Some rare defects in the genes of apoA-I, lecithin cholesterol acyltransferase, and the ATP-binding cassette transporter A1 are the molecular basis of familial HDL deficiency syndromes, but do not contribute much to the variation in HDL cholesterol on the population level (57). Only polymorphisms in the genes of the cholesteryl ester transfer protein, hepatic lipase, and lipoprotein lipase contribute to a similar degree to the variation in HDL cholesterol at the population level as the CAG repeat polymorphism in the AR gene (58 – 60). Previous data on the effect of T on the endothelium are complex and controversial. T was shown to induce and inhibit vasodilatation by endothelium-dependent and endothelium-independent mechanisms by genomic and nongenomic effects (17, 61). Vascular effects induced by T could

4872


be partly mediated via local aromatization to E2, but although aromatase has been identified in vascular smooth muscle cells, it could not be demonstrated in endothelial cells (62). Furthermore, T effects could not be blocked by E antagonists in previous studies (17). The positive correlation between the number of CAG repeats in the AR gene and the flow-induced vasodilatation we describe provides strong evidence that T regulates vascular reactivity by genomic (i.e. AR-dependent) and endothelium-dependent effects. In addition, the association of the CAG repeat polymorphism with glyceroltrinitrate-induced vasodilatation points to the regulation of endothelium-independent vasodilation by genomic effects of T. In agreement with our findings, AR have been localized in endothelial and vascular smooth muscle cells (63– 65). Two different, independent, T-induced effects on the vessel can be assumed (61). The endothelium-dependent pathway is likely to be represented by activation of endothelial nitric oxide synthetase, as T effects can be suppressed by pretreatment with n-␻-nitro-l-arginine methyl ester to block NO synthesis (17). A direct effect on vascular smooth muscle cells is likely to be mediated via ATP-sensitive potassium channels, as vascular T effects can be attenuated by the potassium channel blocker glibenclamide (17). This pathway seems to gain relevance when high concentrations of T are involved (9, 13). These results are consistent with our findings concerning T-mediated influences on endotheliumdependent and -independent vasodilatation. Transmigration of leukocytes into the arterial wall is a critical step in the development of acute and chronic inflammatory lesions in the artery and hence in the pathogenesis of atherosclerosis. This process is mediated by selectins such as E-selectins and subsequently by irreversible binding to adhesion molecules such as VCAM-1. Although VCAM-1 expression was previously reported to be regulated by T (66), we did not find any association of the CAG repeat polymorphism with serum levels of sVCAM-1. Shedding of selectins and adhesion molecules from the surface of the endothelium and the appearance of soluble isoforms in plasma (e.g. sEselectin and sVCAM-1) is known to occur after endothelial injury (67). Thus, the lack of association does not contradict a regulatory effect of T on the expression of adhesion molecules, but mitigates the possibility that the association of the CAG repeat polymorphism with endothelial vasoreactivity is secondary to vascular injury. In conclusion, we demonstrate in a population of healthy men not taking medication of any kind that the CAG repeat sequence polymorphism in the AR gene modulates the metabolic and cardiovascular effects of T. In consequence, a low number of CAG repeats represents a higher risk for reduced endothelial response to ischemia and lower levels of HDL cholesterol. Our results could explain the finding of significantly lower FMD in African Americans compared with Caucasian Americans (68), as the latter seem to have a significantly higher number of CAG repeats (21). Both low HDL cholesterol and endothelial dysfunction of the brachial artery are important risk factors for atherosclerosis (38, 47, 57). Therefore, a low number of CAG repeats in the AR gene appears to put men at increased risk for developing CHD.


Acknowledgments We thank Nicole Terwort and Gabriele Klapdor for excellent technical assistance, and Susan Nieschlag, M.A., for language editing of the manuscript. Received February 22, 2001. Accepted June 25, 2001. Address all correspondence and requests for reprints to: Prof. Dr. E. Nieschlag, Institute of Reproductive Medicine, University of Munster, Domagkstrasse 11, D-48129 Mu¨ nster, Germany. E-mail: [email protected]. This work was supported by a grant from Interdisziplina¨ res Zentrum fu¨ r klinische Forschung Mu¨ nster (Project A3) and the Deutsche Forschungsgemeinschaft Confocal Research Group The Male Gamete: Production, Maturation, Function.

References 1. Alexandersen P, Haarbo J, Christiansen C 1996 The relationship of natural androgens to coronary heart disease in males: a review. Atherosclerosis 125: 1–13 2. Phillips GB, Pinkernell BH, Jing TY 1994 The association of hypotestosteronemia with coronary artery disease in men. Arterioscler Thromb 14:701–706 3. Von Eckardstein A 1998 Androgens, cardiovascular risk factors, and atherosclerosis. In: Nieschlag E, Behre HM, eds. Testosterone: action, deficiency, substitution, 2nd Ed. Berlin, New York: Springer Verlag; 229 –258 4. Zitzmann M, Nieschlag E 2001 Testosterone levels in healthy men and their relation to behavioural and physical characteristics: facts and constructs. Eur J Endocrinol 144:183–197 5. Sullivan ML, Martinez CM, Gennis P, Gallagher EJ 1998 The cardiac toxicity of anabolic steroids. Prog Cardiovasc Dis 41:1–15 6. Bagatell CJ, Bremner WJ 1996 Androgens in men: uses and abuses. N Engl J Med 334:107–114 7. Rosano GMC, Leonardo F, Pagnotta P, et al. 1999 Acute anti-ischemic effect of testosterone in men with coronary artery disease. Circulation 99:1666 –1670 8. Webb CM, Adamson DL, de Zeigler D, Collins P 1999 Effect of acute testosterone on myocardial ischemia in men with coronary artery disease. Am J Cardiol 83:437– 439 9. Webb CM, McNeill JG, Hayward CS, de Zeigler D, Collins P 1999 Effects of testosterone on coronary vasomotor regulation in men with coronary heart disease. Circulation 100:1690 –1696 10. Herman SM, Robinson JT, McCredie RJ, Adams MR, Boyer MJ, Celermajer DS 1997 Androgen deprivation is associated with enhanced endothelium-dependent dilatation in adult men. Arterioscler Thromb Vasc Biol 17:2004 –2009 11. McCrohon JA, Walters WA, Robinson JT, et al. 1997 Arterial reactivity is enhanced in genetic males taking high dose estrogens. J Am Coll Cardiol 29:1432–1436 12. McCredie RJ, McCrohon JA, Turner L, Griffiths KA, Handelsman DJ, Celermajer DS 1998 Vascular reactivity is impaired in genetic females taking high-dose androgens. J Am Coll Cardiol 32:1331–1335 13. Ong PJ, Patrizi G, Chong WC, Webb CM, Hayward CS, Collins P 2000 Testosterone enhances flow-mediated brachial artery reactivity in men with coronary artery disease. Am J Cardiol 85:269 –272 14. Worboys S, Kotsopoulos D, Teede H, McGrath B, Davis S 2001 Evidence that parenteral testosterone therapy may improve endothelium-dependent and -independent vasodilatation in postmenopausal women already receiving estrogen. J Clin Endocrinol Metab 1:158 –161 15. Yue P, Chatterjee K, Beale C, Poole-Wilson PA, Collins P 1995 Testosterone relaxes rabbit coronary arteries and aorta. Circulation 91:1154 –1160 16. Hutchison SJ, Sudhir K, Chou TM, et al. 1997 Testosterone worsens endothelial dysfunction associated with hypercholesterolemia and environmental tobacco smoke exposure in male rabbit aorta. J Am Coll Cardiol 29:800 – 807 17. Chou TM, Sudhir K, Hutchison SJ, et al. 1996 Testosterone induces dilatation of canine coronary conductance and resistance arteries in vivo. Circulation 94:614 –2619 18. Alexandersen P, Haarbo J, Byrjalsen I, Lawetz H, Christiansen C 1999 Natural androgens inhibit male atherosclerosis. A study in castrated, cholesterolfed rabbits. Circ Res 84:813– 819 19. Adams MR, Williams JK, Kaplan JR 1995 Effects of androgens on coronary artery atherosclerosis and atherosclerosis-related impairment of vascular responsiveness. Arterioscler Thromb Vasc Biol 15:562–570 20. Teoh H, Quan A, Leung SW, Man RY 2000 Differential effects of 17␤-estradiol and testosterone on the contractile responses of porcine coronary arteries. Br J Pharmacol 129:1301–1308 21. Platz EA, Rimm EB, Willett WC, Kantoff PW, Giovannucci E 2000 Racial variation in prostate cancer incidence and in hormonal system markers among male health professionals. J Natl Cancer Inst 92:2009 –2017 22. Hsing AW, Gao YT, Wu G, et al. 2000 Polymorphic CAG and GGN repeat


23. 24. 25. 26. 27. 28. 29.

30. 31. 32. 33. 34. 35.

36. 37.

38. 39. 40. 41. 42. 43. 44.

lengths in the androgen receptor gene and prostate cancer risk: a populationbased case-control study in China. Cancer Res 1518:5111–5116 Beilin J, Ball EM, Favaloro JM, Zajac JD 2000 Effect of the androgen receptor CAG repeat polymorphism on transcriptional activity: specificity in prostate and non-prostate cell lines. J Mol Endocrinol 25:85–96 Chamberlain NL, Driver ED, Miesfeld RL 1994 The length and location of CAG trinucleotide repeats in the androgen receptor N-terminal domain affect transactivation function. Nucleic Acids Res 11:3181–3186 Nam RK, Elhaji Y, Krahn MD, et al. 2000 Significance of the CAG repeat polymorphism of the androgen receptor gene in prostate cancer progression. J Urol 164:567–572 Xue W, Irvine RA, Yu MC, Ross RK, Coetzee GA, Ingles SA 2000 Susceptibility to prostate cancer: interaction between genotypes at the androgen receptor and prostate-specific antigen loci. Cancer Res 60:839 – 841 Giovannucci E, Platz EA, Stampfer MJ, et al. 1999 The CAG repeat within the androgen receptor gene and benign prostatic hyperplasia. Urology 53:121–125 Mitsumori K, Terai A, Oka H, et al. 1999 Androgen receptor CAG repeat length polymorphism in benign prostatic hyperplasia (BPH): correlation with adenoma growth. Prostate 41:253–257 Tut TG, Ghadessy FJ, Trifiro MA, Pinsky L, Yong EL 1997 Long polyglutamine tracts in the androgen receptor are associated with reduced transactivation, impaired sperm production, and male infertility. J Clin Endocrinol Metab 82:3777–3782 Yoshida KI, Yano M, Chiba K, Honda M, Kitahara S 1999 CAG repeat length in the androgen receptor gene is enhanced in patients with idiopathic azoospermia. Urology 54:1078 –1081 Dowsing AT, Yong EL, Clark M, McLachlan RI, de Kretser DM, Trounson AO 1999 Linkage between male infertility and trinucleotide repeat expansion in the androgen-receptor gene. Lancet 354:640 – 643 Yong EL, Lim LS, Wang Q, et al. 2000 Androgen receptor polymorphisms and mutations in male infertility. J Endocrinol Invest 23:573–577 Belsham DD, Yee WC, Greenberg CR, Wrogemann K 1992 Analysis of the CAG repeat region of the androgen receptor gene in a kindred with X-linked spinal and bulbar muscular atrophy. J Neurol Sci 112:133–138 Hsiao PW, Lin DL, Nakao R, Chang C 1999 The linkage of Kennedy’s neuron disease to ARA24, the first identified androgen receptor polyglutamine regionassociated coactivator. J Biol Chem 274:20229 –20234 Mariotti C, Castellotti B, Pareyson D, et al. 2000 Phenotypic manifestations associated with CAG-repeat expansion in the androgen receptor gene in male patients and heterozygous females: a clinical and molecular study of 30 families. Neuromuscul Disord 10:391–397 McGill Jr HC, McMahan CA, Herderick EE, Malcom GT, Tracy RE, Strong JP 2000 Origin of atherosclerosis in childhood and adolescence. Am J Clin Nutr 72(Suppl 5):1307S–1315S McGill HC, McMahan CA, Zieske AW, et al. 2000 Associations of coronary heart disease risk factors with the intermediate lesion of atherosclerosis in youth. The Pathobiological Determinants of Atherosclerosis in Youth (PDAY) Research Group. Arterioscler Thromb Vasc Biol 20:1998 –2004 Anderson TJ, Uehata A, Gerhard MD, et al. 1995 Close relation of endothelial function in the human coronary and peripheral circulations. J Am Coll Cardiol 26:1235–1241 Enos WF, Holmes RH, Beyer J 1953 Coronary disease among United States soldiers killed in action in Korea. Preliminary report. JAMA 152:1090 –1093 Stary HC 1994 Changes in components and structure of atherosclerotic lesions developing from childhood to middle age in coronary arteries. Basic Res Cardiol 89:17–32 Woo KS, Chook P, Leong HC, Huang XS, Celermajer DS 2000 The impact of heavy passive smoking on arterial endothelial function in modernized Chinese. J Am Coll Cardiol 36:1228 –1232 Drexler H, Hornig B 1999 Endothelial dysfunction in human disease. J Mol Cell Cardiol 31:51– 60 Esper RJ, Vilarino J, Cacharron JL, et al. 1999 Impaired endothelial function in patients with rapidly stabilized unstable angina: assessment by noninvasive brachial artery ultrasonography. Clin Cardiol 22:699 –703 Toikka JO, Ahotupa M, Viikari JS, et al. 1999 Constantly low HDL cholesterol concentration relates to endothelial dysfunction and increased in vivo LDLoxidation in healthy young men. Atherosclerosis 147:133–138


45. Raitakari OT, Adams MR, McCredie RJ, Griffiths KA, Celermajer DS 1999 Arterial endothelial dysfunction related to passive smoking is potentially reversible in healthy young adults. Ann Intern Med 130:578 –581 46. Schnell GB, Robertson A, Houston D, Malley L, Anderson TJ 1999 Impaired brachial artery endothelial function is not predicted by elevated triglycerides. J Am Coll Cardiol 33:2038 –2043 47. Celermajer DS, Sorensen KE, Bull C, Robinson J, Deanfield JE 1994 Endothelium-dependent dilation in the systemic arteries of asymptomatic subjects relates to coronary risk factors and their interaction. J Am Coll Cardiol 15: 1468 –1474 48. Rubanyi GM 1991 Endothelium-derived relaxing and contracting factors. J Cell Biochem 46:27–36 49. Ignarro LJ, Cirino G, Casini A, Napoli C 1999 Nitric oxide as a signaling molecule in the vascular system: an overview. J Cardiovasc Pharmacol 34: 879 – 886 50. Celermajer DS, Sorensen KE, Gooch VM, et al. 1992 Non-invasive detection of endothelial dysfunction in children and adults at risk of atherosclerosis. Lancet 340:1111–1115 51. Vermeulen A, Verdonck L, Kaufman JM 1999 A critical evaluation of simple methods for the estimation of free testosterone in serum. J Clin Endocrinol Metab 84:3666 –3672 52. Friedewald WT, Levy RI, Fredrickson DS 1972 Estimation of the concentration of low density lipoprotein in plasma without use of a preparative ultracentrifuge. Clin Chem 18:449 – 456 53. Gudivaka R, Schoeller DA, Kushner RF, Bolt MJ 1999 Single- and multifrequency models for bioelectrical impedance analysis of body water compartments. J Appl Physiol 87:1087–1096 54. Kushner RF, Schoeller DA 1986 Estimation of total body water by bioelectrical impedance analysis. Am J Clin Nutr 44:417– 424 55. Pace HV, Rathbun EN 1945 Studies on body composition, body water and chemically combined nitrogen content in relation to fat content. J Biol Chem 158:685– 691 56. Hiort O, Holterhus PM, Horter T, et al. 2000 Significance of mutations in the androgen receptor gene in males with idiopathic infertility. J Clin Endocrinol Metab 85:2810 –2815 57. Von Eckardstein A, Nofer JR, Assmann G 2001 HDL and coronary heart disease: role of cholesterol efflux and reverse cholesterol transport Arterioscler Thromb Vasc Biol 20:13–27 58. Hokanson JE 1999 Functional variants in the lipoprotein lipase gene and risk cardiovascular disease. Curr Opin Lipidol 10: 393–399 59. Inazu, A, Koizumi, J, Mabuchi, H 2000 Cholesteryl ester transfer protein and atherosclerosis. Curr. Opin. Lipidol 11: 389 –396 60. Cohen JC, Vega GL, Grundy SM 1999 Hepatic lipase: new insights from genetic and metabolic studies. Curr Opin Lipidol 10:259 –267 61. Hutchison SJ, Sudhir K, Chou TM, Chatterjee K 1997 Sex hormones and vascular reactivity. Herz 22:141–150 62. Harada N, Sasano H, Murakami H, Ohkuma T, Nagura H, Takagi Y 1999 Localized expression of aromatase in human vascular tissues. Circ Res 84: 1285–1291 63. McGill HC, Sheridan PJ 1981 Nuclear uptake of sex steroid hormones in the cardiovascular system of the baboon. Circ Res 48:238 –244 64. Lin AL, Gonzalez R, Shain SA 1985 Androgen directs apparent cytoplasmic and nuclear distribution of rat cardiovascular androgen receptors. Arteriosclerosis 5:659 – 667 65. Higashiura K, Mathur RS, Halushka PV 1997 Gender-related differences in androgen regulation of thromboxane A2 receptors in rat aortic smooth-muscle cells. J Cardiovasc Pharmacol 29:311–315 66. McCrohon JA, Jessup W, Handelsman DJ, Celermajer DS 1999 Androgen exposure increases human monocyte adhesion to vascular endothelium and endothelial cell expression of vascular cell adhesion molecule-1. Circulation 99:2317–2322 67. Semaan HB, Gurbel PA, Anderson JL, et al. 2000 Soluble VCAM-1 and E-selectin, but not ICAM-1 discriminate endothelial injury in patients with documented coronary artery disease. Cardiology 93:7–10 68. Perregaux D, Chaudhuri A, Rao S, et al. 2000 Brachial vascular reactivity in blacks. Hypertension 36:866 – 871