Serum lipids and estrogen receptor gene polymorphisms in male-to-female transsexuals: effects of estrogen treatment

www.elsevier.com/locate/ejim

European Journal of Internal Medicine 15 (2004) 231–237

Original article

Serum lipids and estrogen receptor gene polymorphisms in male-to-female

transsexuals: effects of estrogen treatment

Manuel Sosaa,b,*, Esteban Jodarc, Elena Arbeloa, Casimira Domınguezd, Pedro Saavedraa,Armando Torrese, Eduardo Salidoe, J.M. Liminanaa, Marıa Jesus Gomez de Tejadaf,

Diego Hernandezb

aGroup of Investigation on Osteoporosis and Bone Metabolic Diseases, University of Las Palmas de Gran Canaria, Las Palmas, SpainbBone Metabolic Unit, Department of Internal Medicine, University Insular Hospital, Las Palmas, Spain

cBone Metabolism Unit, Endocrinology Service, University Hospital 12 de Octubre, Madrid, SpaindDepartment of Biochemistry, Hospital Dr. Negrın, Las Palmas, Spain

e Investigation Unit, Hospital University, University of La Laguna, La Laguna, Tenerife, Canary Islands, SpainfBone Metabolic Unit, Department of Medicine, University of Seville, Seville, Spain

Received 14 October 2003; received in revised form 2 March 2004; accepted 18 March 2004

Abstract

The effects of chronic administration of estrogens on the lipid profile in males are not fully understood. We have studied the effect of

chronic administration of estrogens on the lipid profile in a group of transsexual (TS) Canarian men who were taking estrogens and anti-

androgens for a minimum of 3 years. In this cross-sectional study of cases (n = 27) and controls (n = 26), plasma lipid profile and selected

biochemical and hormonal features were studied. TS subjects had shorter stature than controls, and, after adjusting for height and weight, we

found that they had lower values of serum free testosterone (FT) and higher estradiol (E2) levels than controls. The TS group had lower total

and low-density lipoprotein (LDL) cholesterol and lower apoprotein B (Apo B) levels than the control group. Biochemistry was similar in

both groups. The distribution of estrogen receptor gene polymorphisms (ER-Pvu and ER-Xba) was also similar in both groups. Serum Apo B

concentration was related to ER-Xba polymorphism. No other association between lipid profile and the distribution of ER-Pvu and ER-Xba

was found. We conclude that the chronic administration of estrogens in men could produce an increase in serum estradiol, a decrease in free

testosterone levels, and a reduction in total cholesterol, LDL-cholesterol, and Apo B levels. The ER-Xba polymorphism may influence the

Apo B response to exogenous estrogen in males.

D 2004 Elsevier B.V. All rights reserved.

Keywords: Lipids; Transsexuals; Estrogen receptor gene polymorphism; Estrogens

1. Introduction performed in postmenopausal women. In both the heart and

Estrogen raises plasma levels of high-density lipoprotein

(HDL) cholesterol, an effect that has been related in the past

to the lower rates of coronary heart disease (CHD) in

premenopausal women and in postmenopausal women re-

ceiving estrogen replacement therapy [1]. This perception

recently changed after the publication of two large studies

0953-6205/$ - see front matter D 2004 Elsevier B.V. All rights reserved.

doi:10.1016/j.ejim.2004.04.009

* Corresponding author. Department of Clinical and Surgical

Sciences, Bone Metabolic Unit, Centre of Health Sciences, University

of Las Palmas de Gran Canaria, Apartado 550, 35080 Las Palmas de

Gran Canaria, Las Palmas, Canary Islands, Spain. Tel.: +34-928-451-456;

fax: +34-928-451-428.

E-mail address: [email protected] (M. Sosa).

estrogen/progestin replacement study (HERS) [2] and the

Women’s Health Initiative (WHI) [3,4] studies, the changes

found in lipid levels with hormone therapy were not predic-

tive of CHD outcomes in women with heart disease. Estro-

gen plus progestin did not confer cardiac protection; rather,

the combination increased the risk of CHD among generally

healthy postmenopausal women, especially during the first

year after the initiation of hormone use. Because of this, the

authors concluded that this treatment should not be pre-

scribed for the prevention of cardiovascular disease.

Nevertheless, there is considerable variation in the re-

sponse of HDL-cholesterol levels to endogenous or exoge-

nous estrogens, and among postmenopausal women with

prevalent coronary artery disease, several estrogen receptor

M. Sosa et al. / European Journal of Internal Medicine 15 (2004) 231–237232

gene polymorphisms have shown an augmented response of

HDL-cholesterol to hormone replacement therapy [5].

In healthy men too, estrogen levels could be a predictor of

plasma HDL-cholesterol [6]. HDL-cholesterol concentra-

tions decrease during adolescence in males in association

with increasing pubertal maturation and free testosterone via

decreased Apo AI and Apo AII levels [7]. Long-term

testosterone replacement in hypogonadal men is associated

with a decrease in plasma HDL-cholesterol and Apo AI

concentrations in some studies [8], although there are nu-

merous reports that endogenous testosterone in men, partic-

ularly in aging, is associated with HDL-cholesterol. Indeed,

testosterone can be converted into estrogen in vivo, which

makes it difficult to separate the effects of either hormone.

Nevertheless, the effects of exogenous estrogens in males

are not fully understood. Some studies have been done in

transsexuals (TS) treated with estrogens for prostate cancer.

Non-gonadectomized, male-to-female transsexuals are an-

other good model to study the effect of estrogen and anti-

androgen treatment in male subjects, although gonadecto-

mized male-to-female transsexuals, in whom the results are

not confounded by the use of anti-androgen, are probably

even better.

The purpose of this study was to examine the lipid profile

in a population of non-gonadectomized, male-to-female

transsexuals treated with estrogens and anti-androgens for

at least 3 years. We also looked at the potential role of the

estrogen receptor gene polymorphisms (ER-Pvu and ER-

Xba, also known as estrogen receptor 1, ESR1) in the

response to this therapy.

2. Materials and methods

2.1. Study subjects

Fifty-three men, aged 30–59 years, were enrolled in the

study between December 1999 and July 2000 in Gran

Canaria, Canary Islands (Spain), and divided into two groups.

Group 1 consisted of 27 non-gonadectomized, male-to-

female transsexuals who were on estrogen therapy for

201F108 months (range 3–35 years). The most common

formulations used were contraceptive pills (ethynyl estra-

diol + cyproterone acetate or levonorgestrel), oral estrogens

(conjugated equine), and depot estrogens (estradiol valerate

or mestranol + norethisterone). The doses were mostly in the

pharmacological, rather than in the physiological, range.

Most subjects had used more than one form of therapy, with

frequent use of anti-androgens (cyproterone acetate). Trans-

sexuals were identified through a network of contacts.

Group 2 consisted of 26 healthy men without any intake

of estrogens (control group). They were enrolled in a

random manner as part of other epidemiological studies

performed in our unit.

All patients were Caucasian. No patient was taking

calcium supplements, vitamin D preparations, or other

medications that might affect bone mineral density. None

of the patients had a history of alcoholism, Paget’s disease,

metabolic bone disease, hepatic or renal disorders, or any

other major medical condition. TS patients with a history of

chronic hepatitis, HIV infection, or gonadectomy were

excluded. All patients were informed about the nature of

the study and gave written informed consent. The ethical

review committee of the Hospital University Insular ap-

proved the study, which was conducted in accordance with

the guidelines proposed in the Declaration of Helsinki.

2.2. Questionnaire and physical examination

A previously standardized questionnaire concerning es-

trogen intake and previous diseases and medications that

influence bone density was completed [9]. A complete

physical examination was carried out. Breast development

was graded according to Tanner’s staging. Body mass index

(BMI) was obtained from the equation BMI = body weight/

height2 (kg/m2).

2.3. Sample collection and analytical methods

Serum and urine specimens were obtained after an

overnight fast. Blood was collected without any additives

between 8:00 and 9:00 a.m. After centrifugation at 1500� g

for 10 min, serum was aliquoted and frozen at � 20 jCwithin 1 h of phlebotomy until the biochemical analyses

were performed. Urine samples were collected and stored at

� 20 jC until analysis. Urine creatinine concentrations were

measured with an automated colorimetric method. Serum

glucose, urea, creatinine, inorganic phosphorus, total calci-

um, and alkaline phosphatase (ALP) were determined on a

dry slide chemistry analyzer (Kodak Ektachem Analyzer,

Rochester, NY). Serum total cholesterol and triglyceride

concentrations were also determined by Kodak Ektachem

dry chemistry. Serum HDL-cholesterol was measured with

the use of heparin-manganese precipitation [10]. Low-den-

sity lipoprotein (LDL) cholesterol concentration was calcu-

lated using the Friedewald equation [11]. Apolipoprotein A-

1 (Apo A-1) and apoprotein B (Apo B) were determined by

immunoturbidimetry using a Cobas-Fara II clinical analyzer

(Montclair, NJ, USA). The intra- and interassay coefficients

of variation (CVs) were 1.5% and 2.1%, 2.0% and 2.6%,

1.2% and 2.9%, 1.4% and 3.4%, and 1.4% and 5.2%, for

total cholesterol, triglycerides, HDL-cholesterol, Apo A-1,

and Apo B, respectively.

2.3.1. Hormones

RIA kits were used to analyze serum luteinizing hormone

(LH) and follicle-stimulating hormone [(FSH), CIS Bio

International, France]. The lower limit of the assay is 0.15

mIU/ml for LH and 0.1 mIU/ml for FSH. Serum free

testosterone (FT) and estradiol (E2) were determined by

RIA coat-a-count (Diagnostic Product, USA). The lower

detection limit was 0.15 pg/ml for testosterone and 20 pg/ml

Table 1

General characteristics of the subjects studied (meanF S.D.) and

distribution of certain risk factors for osteoporosis

Transsexuals Controls p

Age (years) 43.0F 7.7 44.0F 6.0 0.6

Height (cm) 169.0F 6.3 174.1F 7.3 0.01

Weight (kg) 74.3F 13.8 81.3F 12.9 0.05

Body mass

index

(kg/m2)

26.0F 4.7 26.7F 3.7 0.5

Actual

calcium

intake

(mg/day)

773.9F 257.9 652.1F 265.6 0.12

Odds ratio; CI 95%

Physical activity during leisure time

Active 36% 48% 0.609; 0.345, 1.073

Sedentary 64% 52%

Tobacco

Yes 48% 40% 1.384; 0.790, 2.424

No 52% 60%

Alcohol consumption

Yes 68% 72% 0.826; 0.450, 1.514

No 32% 28%

Drug consumption

Yes 37.5% 0% NP

No 62.5% 100%

NP: not performed because none of the controls admitted taking drugs.

M. Sosa et al. / European Journal of Inte

for estradiol. The intra- and interassay CVs were 3.1% and

5.9%, 2.7% and 6.8%, 4.0% and 5.5%, and 6.5% and 9.7%

for LH, FSH, FT, and E2, respectively.

Table 2

Biochemical values, lipid levels, and hormone levels in the groups studied

Transsexuals Con

Biochemistry

Urea (mg/dl) 34.8F 8.6 38

Creatinine (mg/dl) 0.8F 0.2 0

Glucose (mg/dl) 97.2F 15.5 99

Calcium (mg/dl) 9.1F 0.4 9

Phosphate (mg/dl) 3.0F 0.5 3

Uric acid (mg/dl) 4.7F 1.1 5

Total proteins (g/l) 7.228F 0.349 7.34

Lipids

Total cholesterol (mg/dl) 184.2F 40.7 222

Triglycerides (mg/dl) 114.5F 55.4 132

HDL-cholesterol (mg/dl) 46.0F 14.3 49

LDL-cholesterol (mg/dl) 119.4F 38.6 144

Apoprotein A (mg/dl) 138.6F 69.7 147

Apoprotein B (mg/dl) 85.8F 28.2 106

Hormones

LH (mIU/ml) 4.5F 3.3 3

FSH (mIU/ml) 6.0F 4.4 5

Free testosterone (pg/ml) 9.6F 9.3 17

Estradiol (pg/ml) 175.9F 252.8 42

* p of comparisons adjusted for weight and height.

2.3.2. Polymorphisms of the estrogen receptor gene

DNA was purified from 3 ml of blood containing EDTA

using proteinase K digestion, phenol extraction, and ethanol

precipitation. To determine the polymorphic sequences in

the first intron of ER, a 1300-bp DNA fragment was

amplified with primers ER-1: 5V-CTGCCACCCTATCTG-TATC-3V and ER-2: 5V-ACCCTGGCGTCGATTATCTG-3V at 94 jC (1 min), 56 jC (1 min), and 72 jC (2 min) at

30 cycles. Ten microliters of the amplification product was

digested with either 5 U PvuII or 5 U XbaI (New England

Biolabs) at 37 jC for 3 h and products were resolved by

electrophoresis in 2% agarose gels. The ER p allele (PvuII

digestion) was observed as two bands of 1000 and 300 bp,

while the P allele remained as a single 1300-bp band.

Similarly, the ER x allele (XbaI digestion) was observed

as two bands of 900 and 400 bp, while the X allele remained

as a single 1300-bp band.

2.4. Statistical analysis

All results were expressed as meanF S.D. unless other-

wise indicated. A p-level below 0.05 was considered sig-

nificant. First we applied the Kolmogorov–Smirnov test to

ascertain the normal distribution of the parameters studied.

Then, we compared general characteristics of TS and con-

trols using Student’s t-test. For categorical variables, we

applied the chi-square test and calculated the odds ratios. To

make comparisons between multiple groups, we used the

Scheffe and Student’s–Neuman–Keuls tests. All of the

comparisons were made after regression adjustment by

weight and height. Finally, we performed the Pearson

correlation test to study the correlations between lipid

rnal Medicine 15 (2004) 231–237 233

trols p Unadjusted p Adjusted*

.5F 19.2 0.01 0.03

.9F 0.1 0.009 0.08

.2F 14.7 0.6 0.5

.3F 0.4 0.4 0.6

.3F 0.4 0.2 0.3

.3F 1.5

4F 0.451 0.3 0.2

.4F 32.5 0.002 0.05

.6F 104.2 0.4 0.3

.1F 9.7 0.4 0.6

.0F 28.7 0.04 0.03

.6F 22.8 0.6 0.1

.3F 21.9 0.05 0.05

.7F 1.5 0.3 0.3

.4F 2.0 0.6 0.1

.5F 4.2 0.01 0.001

.0F 13.2 0.02 0.01

Table 3

Polymorphisms of VDR and ER-Pvu/ER-Xba: analysis of gene frequency

Transsexuals (%) Controls (%) p

ER-Pvu-1

PP 14.8 16.0 0.306

Pp 44.4 52.0

pp 40.7 32.0

ER-Xba

XX 3.7 12.0 0.296

Xx 48.1 48.0

xx 48.1 40.0

M. Sosa et al. / European Journal of Internal Medicine 15 (2004) 231–237234

fractions and hormones, both in TS and in controls. All

studies were performed using SPSS, version 11.0 (SPSS, IL,

USA).

3. Results

Table 1 shows the general characteristics of both groups.

TS subjects were shorter and lighter in weight than controls.

Because of this, all of the statistical studies performed were

done after adjusting for these two parameters. Nevertheless,

the BMI was comparable in both groups. No significant

differences were found between the TS and control groups

with regard to physical activity during leisure time or to

tobacco and alcohol consumption. Indeed, the actual calci-

um intake was similar in both groups. Some 37.5% of TS

subjects had a history of regular drug consumption, mainly

cannabis and cocaine, while none of the control subjects

admitted having this habit. Gynecomastia equivalent to

Table 4

Comparison of lipid and hormonal values between transsexuals and controls, dep

Variable Polymorphisms

pp pP/Pp

Transsexuals

Total cholesterol (mg/dl) 175.3F 37.8 200.1FHDL-cholesterol (mg/dl) 42.5F 10.3 50.6FLDL-cholesterol (mg/dl) 107.6F 22.8 137.4FApoprotein A (mg/dl) 138.5F 55.5 137.2FApoprotein B (mg/dl) 102.9F 19.4 78.1FTriglycerides (mg/dl) 96.8F 47.5 130.4FFree testosterone (pg/ml) 11.5F 10.1 8.8FEstradiol (pg/ml) 139.7F 55.9 168.8F

Controls

Total cholesterol (mg/dl) 209.2F 25.2 232.8FHDL-cholesterol (mg/dl) 56.0F 9.3 45.2FLDL-cholesterol (mg/dl) 140.0F 20.6 152.7FApoprotein A (mg/dl) 159.1F 20.8 139.5FApoprotein B (mg/dl) 95.3F 12.1 118.1FTriglycerides (mg/dl) 66.0F 19.3 151.5FFree testosterone (pg/ml) 16.2F 3.3 17.5FEstradiol (pg/ml) 40.0F 13.7 40.7F

a SNK: Student’s–Newman–Keuls test.

Tanner II–III was present in all TS subjects without breast

implants.

Biochemical data, including hormone levels, are pre-

sented in Table 2. Not surprisingly, TS subjects had lower

free testosterone levels and higher estradiol levels than

controls. Serum urea and creatinine were also lower in TS

subjects. When biochemical data were adjusted for weight

and height, the differences between the TS and control

groups with regard to free testosterone, estradiol, and urea

were maintained, while the difference in creatinine levels

disappeared. Plasma lipids showed a more favorable profile

in TS subjects after adjusting for weight and height, with

lower total and LDL-cholesterol and Apo B lipoprotein

concentrations.

The prevalence of the ER-Pvu and ER-Xba polymor-

phisms in our population is shown in Table 3. As expected,

no significant differences in the distribution of ER-Pvu/ER-

Xba polymorphisms were found between patients and con-

trols. There was no segregation of the phenotype with any of

the genotypes.

Table 4 shows the comparison of lipid values in TS and

controls, depending on the Pvu estrogen receptor polymor-

phism. To make a comparison between several groups, we

applied both the Student’s–Newman–Keuls test and the

Scheffe tests. We found no statistical differences between

total cholesterol, its subfractions, triglycerides, Apo A, Apo

B, estradiol, or free testosterone in TS and controls and their

Pvu estrogen receptor polymorphism, with the only excep-

tion of triglycerides in controls.

Table 5 shows the comparison of lipid and hormonal

values between TS and controls, depending on the Xba

estrogen receptor polymorphism. We found only three

ending on the Pvu estrogen receptor polymorphism

p values adjusted

PP(SNKa/Scheffe)

31.7 157.3F 64.5 0.211/0.240

17.4 39.6F 8.7 0.504/0.535

38.6 97.3F 52.1 0.278/0.308

98.5 142.5F 21.9 0.996/0.996

30.6 67.3F 33.6 0.300/0.329

64.5 103.0F 33.0 0.607/0.634

9.6 8.2F 8.1 0.242/0.272

98.4 197.9F 69.6 0.138/0.161

37.0 219.7F 29.8 0.431/0.463

8.4 45.7F 8.3 0.114/0.135

32.8 133.7F 34.5 0.525/0.555

22.4 132.0F 23.6 0.564/0.545

17.1 88.4F 15.4 0.433/0.521

90.8 202.0F 170.0 0.05/0.07

3.4 19.5F 7.1 0.327/0.360

10.4 50.8F 20.4 0.341/0.374

Table 5

Comparison of lipid and hormonal values between transsexuals and

controls, depending on the Xba estrogen receptor polymorphism

Variable Polymorphism p-value

xx xX and XX

Transsexuals

Total cholesterol (mg/dl) 175.0F 35.4 190.3F 44.3 0.425

HDL-cholesterol (mg/dl) 41.3F 9.5 48.8F 16.3 0.329

LDL-cholesterol (mg/dl) 112.5F 23.6 124.1F 46.9 0.587

Apoprotein A (mg/dl) 137.6F 48.1 139.5F 88.6 0.966

Apoprotein B (mg/dl) 105.3F 17.7 66.4F 23.1 0.017

Triglycerides (mg/dl) 95.0F 44.3 127.5F 59.9 0.206

Free testosterone (pg/ml) 9.3F 10.1 8.8F 9.6 0.916

Estradiol (pg/ml) 128.1F 45.1 261.2F 115.8 0.281

Controls

Total cholesterol (mg/dl) 209.3F 34.5 239.4F 25.8 0.052

HDL-cholesterol (mg/dl) 52.2F 11.0 48.3F 8.9 0.448

LDL-cholesterol (mg/dl) 132.7F 28.0 154.6F 28.4 0.443

Apoprotein A (mg/dl) 151.8F 23.5 143.8F 24.0 0.531

Apoprotein B (mg/dl) 98.0F 20.5 121.7F 17.1 0.037

Triglycerides (mg/dl) 99.7F 70.8 170.6F 138.5 0.191

Free testosterone (pg/ml) 15.9F 3.7 18.0F 2.9 0.459

Estradiol (pg/ml) 38.4F 12.8 40.9F 11.0 0.651

M. Sosa et al. / European Journal of Internal Medicine 15 (2004) 231–237 235

controls and no TS with the polymorphism XX, so we

decided to group them all into two groups, group I had only

the xx polymorphism and group II had xX and XX poly-

morphisms. The xx polymorphism was associated with

higher Apo B levels in control males and with lower Apo

B levels in male-to-female TS.

Table 6 shows the correlation between lipids and hor-

mones in TS. We found positive correlations between Apo

A and HDL-cholesterol and between Apo A and serum

estradiol and a statistically significant negative correlation

between estradiol and LDL-cholesterol.

Finally, in Table 7, we present the correlation between

lipids and hormones in controls. In this group, we found no

statistically significant correlations between estradiol and

lipids, and the correlations found were between lipids. Thus,

Table 6

Correlations between lipids and hormones in group 1 (transsexuals)

Total cholesterol HDL-cholesterol LDL-cholest

Total cholesterol –

HDL-cholesterol r = 0.424 –

p= 0.340

LDL-cholesterol r = 0.728 r =� 0.294 –

p= 0.06 p= 0.52

Triglycerides r = 0.588 r = 0.523 r = 0.129

p= 0.16 p= 0.22 p= 0.782

Apo A r = 0.224 r = 0.942 r =� 0.459

p= 0.620 p= 0.001 p= 0.30

Apo B r = 0.706 r = 0.140 r = 0.666

p= 0.07 p= 0.763 p= 0.10

Estradiol r =� 0.340 r = 0.635 r =� 0.865

p= 0.45 p= 0.125 p= 0.01

Free testosterone r = 0.202 r =� 0.364 r = 0.468

p= 0.66 p= 0.42 p= 0.28

we found a very good correlation between total cholesterol

and LDL-cholesterol, between HDL-cholesterol and trigly-

cerides, between HDL-cholesterol and Apo A, between total

cholesterol and Apo B, between LDL-cholesterol and Apo

B, and, finally, between triglycerides and Apo A.

4. Discussion

The effect of oral contraceptives on serum lipid values in

normal women depends upon the estrogen dose and the

androgenicity of the progestin. In general, serum triglycer-

ide concentrations increase slightly, but there are no consis-

tent changes in serum HDL or LDL lipoprotein cholesterol

concentrations. The estrogen component of oral contracep-

tives increases serum triglycerides and HDL concentrations

and lowers serum LDL-cholesterol concentrations. These

potentially favorable effects, which could contribute to a

beneficial effect of estrogen on cardiovascular risk [12,13],

have been under discussion since the publication of the

results of the HERS and WHI studies in postmenopausal

women [2–4]. It was a common belief that estrogen

replacement therapy in postmenopausal women had a fa-

vorable effect on coronary risk by changing the lipid profile,

i.e., mainly reductions in LDL-cholesterol (15%) and lipo-

protein A (20%) and elevations in HDL-cholesterol (10–

15%) and triglycerides (24%) [1,14]. However, the unfa-

vorable results found have since produced profound changes

in the use of hormone replacement therapy [15].

In healthy men, estrogen levels are also the main pre-

dictors of plasma HDL-cholesterol [6], with puberty being

associated with a decrease in HDL-cholesterol [7] and the

hypogonadal state associated with an increase in plasma

HDL-cholesterol and Apo AI concentrations [8]. Lipopro-

tein analysis from a male subject with estrogen insensitivity

syndrome showed relatively low levels of total, LDL-, and

HDL-cholesterol, Apo AI, and lipoprotein A, but normal

levels of triglycerides and pre-beta-1-HDL cholesterol.

erol Triglycerides Apo A Apo B Estradiol

–

r = 0.321 –

p= 0.48

r =� 0.384 r = 0.153 –

p= 0.93 p= 0.74

r = 0.233 r = 0.771 r =� 0.419 –

p= 0.61 p= 0.04 p= 0.34

r =� 0.072 r =� 0.509 r = 0.257 r =� 0.681

p= 0.87 p= 0.242 p= 0.57 p = 0.09

Table 7

Correlations between lipids and hormones in group 2 (controls)

Total cholesterol HDL-cholesterol LDL-cholesterol Triglycerides Apo A Apo B Estradiol

Total cholesterol –

HDL-cholesterol r = 0.295 –

p= 0.32

LDL-cholesterol r = 0.959 r = 0.210 –

p= 0.000 p= 0.49

Triglycerides r = 0.157 r =� 0.616 r = 0.012 –

p= 0.60 p= 0.02 p= 0.96

Apo A r = 0.395 r = 0.818 r =� 0.055 r =� 0.508 –

p= 0.89 p= 0.001 p= 0.85 p= 0.07

Apo B r = 0.830 r =� 0.112 r = 0.806 r = 0.436 r =� 0.107 –

p= 0.000 p= 0.71 p= 0.001 p= 0.136 p= 0.72

Estradiol r = 0.159 r = 0.056 r = 0.144 r = 0.026 r =� 0.283 r =� 0.132 –

p= 0.603 p= 0.85 p= 0.63 p= 0.931 p= 0.34 p= 0.66

Free testosterone r =� 0.087 r =� 0.051 r =� 0.101 r = 0.052 r = 0.215 r = 0.192 r = 0.007

p= 0.77 p= 0.86 p= 0.74 p= 0.86 p= 0.480 p= 0.52 p= 0.98

M. Sosa et al. / European Journal of Internal Medicine 15 (2004) 231–237236

More interestingly, premature coronary artery disease was

discovered in this male patient with a disruptive mutation in

the estrogen receptor gene [16]. In fact, animal models have

shown that the atheroprotective effects of estradiol are

mediated mainly by ER-1 [17].

The administration of tamoxifen to healthy boys with

pubertal gynecomastia produced moderate, but significant,

decreases in total cholesterol and lipoprotein(a) [18]. Yet,

previous studies on the effects of estrogens on lipids in

males are scarce. Estrogen reduced serum levels of trigly-

cerides and LDL-cholesterol in one hypercholesterolemic

male with very low LDL-receptor activity, at least in part,

via an increase in LDL-receptor activity [19].

Non-gonadectomized, male-to-female TS represent a

good model for studying the effects of estrogens in males.

We have shown that these individuals have lower total and

LDL-cholesterol levels than would be expected of males of

their age after treatment with estrogens and anti-androgens

for 3 years or longer. This more favorable lipid profile was

associated with reduced free testosterone and increased

estrogen levels, and it was unrelated to the distribution of

ER-Pvu/ER-Xba polymorphisms.

At this point, we would like to comment on some

limitations of our study. One is that ethynyl estradiol, used

by most of the participants, cannot be measured with a

standard estradiol assay. Also, the small number of subjects

and the heterogeneity of the estrogen preparations used

make it difficult to establish a definitive relationship be-

tween genotype of the estrogen receptor and lipid profiles.

Numerous naturally occurring polymorphisms of human

ER-1 are associated with several clinical phenotypes, includ-

ing risk of breast cancer [20], risk of abortion [21], bone

mineral density [22], anti-fracture efficacy of hormone

replacement therapy [23], body mass index [24], hyperten-

sion [25], coronary atherosclerosis [26], serum lipids [26],

and even augmented response of HDL to estrogen replace-

ment therapy [5].

Matsubara et al. [26] did not find any relationship

between ER polymorphisms and serum lipids in men. Our

data in male-to-female, estrogen-treated TS showed an

interaction between Apo B levels and ER polymorphisms.

The different forms, dosages, and consistency of estro-

gens used may have affected the results of this study. This is

an unavoidable limitation that we were aware of as soon as

we began to collect the data because every TS took estrogens

and anti-androgens in their own way. We conclude that

chronic administration of estrogens in men may produce a

better lipid profile by reducing both total cholesterol and its

subfraction of LDL-cholesterol. This study shows that plas-

ma lipids in adult men are probably sensitive to estrogens.

Although the obvious untoward effects of estrogens in males

preclude their use for hyperlipidemia, these findings corrob-

orate the notion that activation of the ER results in favorable

effects on serum lipids in males. They also suggest a

potential role for ER polymorphisms in the modulation of

this response, thus strengthening the rationale for developing

selective estrogen receptor modulators for use in men.

Acknowledgements

This work was supported by a grant from the College of

Physicians of Las Palmas and from Gerican SL, together

with the University Foundation of Las Palmas de Gran

Canaria.

References

[1] Sullivan JM, Vander Zwaag R, Lemp GF, Hughes JP, Maddock V,

Kroetz FW, et al. Postmenopausal estrogen use and coronary athero-

sclerosis. Ann Intern Med 1988;108:358–63.

[2] Shlipak MG, et al., Heart and Estrogen/Progestin Replacement Study

Investigators. Lipid changes on hormone therapy and coronary heart

disease events in the heart and estrogen/progestin replacement study

(HERS). Am Heart J 2003;146:870–5.

[3] Writing Group for the Women’s Health Initiative Investigators. Risks

and benefits of estrogen plus progestin in healthy postmenopausal

women. JAMA 2002;288:321–3.

[4] Manson JE, Hsia J, Johnson KC, Rossouw JE, Assaf AR, Lasser NL,

M. Sosa et al. / European Journal of Internal Medicine 15 (2004) 231–237 237

et al. Estrogen plus progestin and the risk of coronary heart disease.

N Engl J Med 2003;349:523–34.

[5] Herrington DM, Howard TD, Hawwkins GA, Reboussin DM, Xu J,

Zheng SL, et al. Estrogen receptor polymorphisms and effects of

estrogen replacement on high-density lipoprotein cholesterol in wom-

en with coronary artery disease. N Engl J Med 2002;346:967–74.

[6] Vatalas IA, Dionyssious-Asteriou A. Adrenal C19 steroids and serum

lipoprotein levels in healthy men. Nutr Metab Cardiovasc Dis 2001;

11:388–93.

[7] Morrison JA, Sprecher DL, Biro FM, Apperson-Hausen C, Dipaola

LM. Serum testosterone associates with lower high-density lipopro-

tein cholesterol in black and whites males, 10 to 15 years of age,

through lowered apolipoprotein AI and AII concentrations. Metabo-

lism 2002;51:432–7.

[8] Berg G, Schreier L, Geloso G, Otero P, Nagelberg A, Levalle O.

Impact on lipoprotein profile after long-term testosterone replacement

in hypogonadal men. Horm Metab Res 2002;34:87–92.

[9] Grupo de trabajo sobre protocolos y practica clınica SEIOMM. Datos

basicos en osteoporosis. Rev Esp Enferm Metab Oseas 2000;9:84–5.

[10] Burstein M, Samaille J. Sur un dosage rapide du cholesterol lie aux

a-et aux h-lipoproteines du serum. Clin Chim Acta 1960;5:609.

[11] Friedewald WT, Levy RI, Fredickson DS. Estimation of the concen-

tration of low density lipoprotein cholesterol in plasma without use of

the preparative ultracentrifuge. Clin Chem 1972;18:499–502.

[12] Martin KA, Douglas PS. In: Rose BD, editor. Risks and side effects

associated with oral contraceptives; 2003. Up to date, Wellesley, MA.

[13] Lobo RA, Skinner JB, Lippman JS, Cirillo SJ. Plasma lipids and

desogestrel and ethynyl estradiol: a meta-analysis. Fertil Steril 1996;

65:1100–9.

[14] Binder EF, Williams DB, Schechtman KB, Jeffe DB, Kohrt WM.

Effects of hormone replacement therapy on serum lipids in elderly

women. A randomized, placebo-controlled trial. Ann Intern Med

2001;134:754–60.

[15] Laton B, Rose S, McLeod D, Dowell A. Changes in the use of

hormonal replacement therapy after the report from Women’s Health

Initiative: a cross sectional survey of users. BMJ 2003;327:845–6.

[16] Sudhir K, Chou TM, Chatterjee K, Smith EP, Williams TC, Kane JP,

et al. Premature coronary artery disease associated with a disruptive

mutation in the estrogen receptor gene in a man. Circulation 1997;

96:3774–7.

[17] Hodgin JB, Krege JH, Reddick RL, Korach KS, Smithies O, Maeda

N. Estrogen receptor alpha is a major mediator of 17beta-estradiol’s

atheroprotective effects on lesion size in Apoe� /� ?? mice. J Clin

Invest 2001;107:333–40.

[18] Novoa FJ, Boronat M, Carrillo A, Tapia M, Diaz-Cremades J, Chir-

ino R. Effects of tamoxifen on lipid profile and coagulation param-

eters in male patients with pubertal gynecomastia. Horm Res 2002;

57:187–91.

[19] Inukai T, Takanashi K, Takebayashi K, Tayama K, Aso Y, Takiguchi

Y, et al. Estrogen markedly increases LDL-receptor activity in hyper-

cholesterolemic patients. J Med 2000;31:247–61.

[20] Andersen TI, Heimdal KR, Skerede M, Tveit K, Berg K, Borresen

AL. Oestrogen receptor (ESR) polymorphisms and breast cancer sus-

ceptibility. Hum Genet 1994;94:665–70.

[21] Lehrer S, Sanchez M, Song HK, Dalton J, Levine E, Savoretti P, et al.

Oestrogen receptor B-region polymorphisms and spontaneous abor-

tion in women with breast cancer. Lancet 1990;335:622–4.

[22] Kobayashi S, Inoue S, Hosoi T, Ouchi Y, Shiraki M, Orimo H. As-

sociation of bone mineral density with polymorphisms of the estrogen

receptor gene. J Bone Miner Res 1996;11:306–11.

[23] Salmen T, Heikkinen AM, Mahonen A, Kroger H, Komulainen M,

Saarikoski S, et al. The protective effect of hormone replacement

therapy on fracture risk is modulated by estrogen receptor alpha ge-

notype in early postmenopausal women. J Bone Miner Res 2000;15:

2479–86.

[24] Deng HW, Li J, Li JL, Dowd R, Davies KM, Johnson M, et al.

Association of estrogen receptor-alpha genotypes with body mass

index in normal healthy postmenopausal Caucasian women. J Clin

Endocrinol Metab 2000;85:2748–51.

[25] Lehrer S, Rabin J, Kalir T, Schachter BS. Estrogen receptor variant

and hypertension. Hypertension 1993;21:439–41.

[26] Matsubara Y, Murata M, Kawano K, Zama T, Aoki N, Yoshino M,

et al. Genotype distribution of estrogen receptor polymorphisms in

men and postmenopausal women from healthy and coronary popula-

tions and its relation to serum lipid levels. Arterioscler Thromb Vasc

Biol 1997;17:3006–12.