| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Articles |
Section on Pediatric Endocrinology, Developmental Endocrinology Branch, National Institute of Child Health and Human Development, National Institutes of Health (C.A.S., A.V., L.S.K., C.S.M., G.P.C.), Bethesda, Maryland 20892; the Department of Pharmacology, University of Maryland (A.B., Q.L., W.Y.), Baltimore, Maryland 21201; Southview Medical Group/St. Vincent’s Hospital (D.D.), Birmingham, Alabama 35205; and the Department of Pediatrics, Georgetown University (C.A.S., A.W.F.), Washington, D.C. 20007
Address all correspondence and requests for reprints to: Dr. Constantine A. Stratakis, Unit on Genetics and Endocrinology, Section on Pediatric Endocrinology, Developmental Endocrinology Branch, National Institute of Child Health and Human Development, National Institutes of Health, Building 10, Room 10N 262, 10 Center Drive, MSC 1862, Bethesda, Maryland 20892-1862. E-mail: stratakc{at}cc1.nichd.nih.gov
Abstract
Increased extraglandular aromatization has been reported as the cause of familial gynecomastia. We studied a kindred with aromatase excess inherited in an autosomal dominant manner, in which affected males had heterosexual precocity and/or gynecomastia, and affected females had isosexual precocity and/or macromastia.
The propositus was a 9-yr-old boy with gynecomastia. His
7.5-yr-old sister had precocious puberty, and their father and paternal
grandmother had peripubertal gynecomastia and macromastia,
respectively. Serum concentrations of gonadal and adrenal steroid
hormones were determined before and after the administration of
corticotropin and/or hCG. Aromatase activity was determined by
[3H]
4-androstenedione to
[3H]estrone conversion by cultured skin fibroblasts
and/or Epstein-Barr virus-transformed lymphocytes and was detected by
immunohistochemistry and/or Western analysis. Linkage was examined with
a polymorphism of the aromatase (P450arom) gene. The P450arom messenger
ribonucleic acid was analyzed by rapid amplification of complementary
DNA (cDNA) ends, ribonuclease protection assay, and RT-PCR.
hCG testing demonstrated a high rate of conversion of
4-androstenedione to estrone and of testosterone to
estradiol in the propositus and his father. Treatment of the propositus
and his sister was initiated with an aromatase inhibitor (testolactone)
and a GnRH analog, which successfully delayed skeletal and pubertal
development in both children. Markedly increased aromatase activity was
found in the patients’ fibroblasts and Epstein-Barr virus-transformed
lymphocytes. The P450arom polymorphism segregated with the disease in
the family. A new 5'-splice variant was present in the patients’
P450arom messenger ribonucleic acid, thus identifying yet another first
exon of this gene, which appears to be aberrantly expressed in this
family.
In conclusion, a family with the aromatase excess syndrome is described, in which the condition was inherited in an autosomal dominant manner, led to feminizing manifestations in both sexes, and was associated with the aberrant utilization of a novel transcript of the P450arom gene.
SOME BREAST development occurs in 30–65% of boys during puberty, but is extremely uncommon in the prepubertal years (1, 2, 3). Prepubertal gynecomastia is most frequently caused by exogenously administered estrogens or hormone-secreting tumors (4). Rarely, this condition is inherited as a familial trait (Mendelian inheritance in man 107910), which has been associated with a high rate of conversion of serum C19 to C18 steroids (5, 6). Increased extraglandular aromatization was first reported in a boy with prepubertal gynecomastia in 1977 (7), then in five members of an African-American kindred in 1985 (8). Interestingly, idiopathic pubertal gynecomastia has been associated with an elevated ratio of serum estrogen to androgen levels during early puberty (9), and excessive aromatase activity in cultured pubic skin fibroblasts from these patients with gynecomastia (10).
Aromatase, also known as estrogen synthetase, is the key enzyme in estrogen biosynthesis. It is an enzymatic complex composed by the aromatase cytochrome P450 (P450arom) and the flavoprotein NADPH-P450 reductase, which catalyzes the conversion of androgens to estrogens (11). Cytochrome P450arom, like the other heme-binding P450 isoforms, is a highly conserved, ancient gene, that is widely expressed in a variety of species in a complex, tissue-specific manner. The human P450arom gene contains 10 exons and is located on chromosome 15q21, an area syntenic to mouse chromosome 9, which harbors a cluster of P450 genes (12, 13, 14). It is expressed in the placenta, ovary, testis, brain, skin fibroblasts, adipocytes, normal breast and breast cancer epithelial and stromal cells, and a number of fetal tissues, including liver, brain, and intestine (11). Tissue-specific regulation of the P450arom messenger ribonucleic acid (mRNA) is ensured by alternative splicing of exon 1 and part of exon 2, driven by at least five major promoter regions located in the 5'-untranslated end of the gene (15, 16, 17, 18). A number of additional minor cDNA species have been identified in various tissues, although their functional significance remains uncertain (11, 15). Both sequences with enhancing and silencing function as well as known responsive elements to a variety of transcription factors have been identified in the regulatory region upstream of the P450arom gene (19, 20, 21, 22). Because of its complex regulation of function, the 5'-end of the P450arom gene has long been considered a candidate for disorders in which inappropriate aromatase activity is encountered. Indeed, the 5'-end splicing products of the P450arom gene in mammary neoplastic and human myeloid leukemia cells differ from those in the normal breast or represent new variants, respectively (16, 23, 24). Similarly, the study of uterine, ovarian, and testicular tumors of varying types and grades of malignancy suggested that alternative promoter use was associated with the mechanisms underlying neoplastic transformation (16, 25, 26). Increased aromatization is the mechanism behind the henny feathering trait in an animal model, the Sebright bantam chicken (27, 28).
In the present study, we investigated a kindred with several members expressing clinical and biochemical evidence of aromatase excess. We found increased aromatase activity in cell lines established from the patients and identified a novel exon 1 of the P450arom gene that was uniquely transcribed in the cells from affected individuals.
Subjects and Methods
Subjects
The propositus is a member of a large pedigree (Fig. 1A
). He, his sister, father, and mother
were seen at the NIH Clinical Center Pediatric Endocrinology clinic
under a protocol approved by the institutional review board of the
NICHD for the investigation of pediatric endocrine disorders
(94-CH-102). The proband, currently a 12-yr-old boy, had presented to
one of the authors (D.D.) at the age of 9 yr with gynecomastia and
accelerated growth. At that time, his high resolution peripheral blood
lymphocyte (PBL) karyotype was 46,XY; imaging studies failed to show
any tumors; his bone age was 13.5 yr; he had elevated estradiol
[E2; 18 pg/mL (66.1 pmol/L)] and estrone
[E1; 69 ng/dL (2546 pmol/L)], prepubertal
testosterone [T; 10 ng/dL (0.36 nmol/L)], and undetectable hCG
levels. The patient was first seen at the NIH at the age of 10 yr. His
examination was remarkable for a height and weight appropriate for a
13-yr-old boy, well formed and homogeneous breasts bilaterally at
Tanner pubertal stage IV (TS-IV; Fig. 1B), pubic hair at TS-III, and
testicular volume of 4.5 mL bilaterally. The propositus’ sister, at
age 7.5 yr had height and weight appropriate for 9.5- and 12-yr-old
girls, respectively; breast development and pubic hair at TS-III with
absence of clitoromegaly; and a normal ano-genital ratio (0.35; Fig. 1C). The father of the two children, a 35-yr-old man, had a height of
165.2 cm and a weight of 86.5 kg (3rd and 90th percentile,
respectively). His examination revealed a normally masculinized male
with a relatively small testicular volume (15 mL bilaterally). At age
15 yr, he had undergone bilateral mastectomy for severe gynecomastia,
first noted at age 11 yr, at which time he entered puberty. He had no
impotence or decreased libido; he refused semen evaluation. He and his
wife had no other children. The mother of the children was 37 yr old
and had a normal physical examination, gynecological evaluation, and
history. The paternal grandmother reportedly had macromastia; a
paternal aunt was 31 yr old and healthy, whereas another aunt was 23 yr
old and had menstrual irregularities. Maternal family history was
unremarkable.
|
Magnetic resonance imaging scans of the head, chest, abdomen, and pelvis were obtained, and bone age was determined in both children. The testes of the propositus and the ovaries of his sister were evaluated by sonography. In the two children, baseline and ACTH (250 µg, iv)-stimulated levels of adrenal steroid hormones and their precursors, and baseline and GnRH (100 µg, iv)-stimulated levels of gonadotropins and gonadal steroids were also determined. The propositus and his father also had a short (1500 U, im, and blood sampling at 48 h) and a long (100 U/kg, im, and blood sampling at 12, 24, 48, 72, and 96 h) hCG stimulation test, as described previously (29, 30, 31). Final height prediction was obtained using the Bayley-Pinneau method (32, 33). Both children were started on treatment with an aromatase inhibitor, testolactone (40 mg/kg·day, orally, divided into three daily doses) and a GnRH analog (histrelin acetate, 10–15 µg/kg·dose, as a single sc injection each evening), as previously described (34, 35). For the last 3 yr, the propositus and his sister have been followed with biannual physical examinations, GnRH stimulation tests, bone age evaluations, and determination of serum estrogen levels.
For hormone assays, blood samples were centrifuged immediately after collection, and serum was stored at -20 C until assayed. All hormones were measured by RIA (Hazelton Biotechnologies, Vienna, VA), as previously described (34, 35).
Cell lines, aromatase assay, and immunocytochemistry
PBL were extracted by the Ficoll method from the propositus, his
sister, father, mother, and paternal grandparents; 5 x
106 of these cells were used for permanent lymphoblastoid
cell line establishment by transformation with the Epstein-Barr virus
(EBV), as previously described (36). The remaining cells were frozen in
liquid nitrogen and stored at -70 C. A skin biopsy was performed after
local anesthesia in the dorsal forearm area, and fibroblasts were grown
by standard methods from three individuals: the father, the mother, and
their daughter. Aromatase activity was measured by
4-androstenedione (A) to E1
conversion by primary cultures of skin fibroblasts frozen and regrown
only once, PBLs, and EBV-transformed lymphoblasts (EBV-tL), and
expressed as femtomols of estrogen per mg protein/h, as previously
described (37, 38) The assays were performed twice, and the mean was
calculated. Briefly, cell cultures were washed with Hanks’ solution
before the addition of 1 ml assay medium containing approximately 0.3
µCi 1ß-[3H]A and unlabeled substrate. The culture
plates were then placed in an incubator (5% CO2) at 37 C.
After 2 h, the medium was transferred to a test tube, and 2 mL
chloroform were added. The unconverted substrate and steroid products
were extracted into the organic phase. An aliquot of 0.7 mL of the
aqueous phase was treated with 2.5% activated, dextran-coated charcoal
suspension to remove residual steroids. Tritiated water
(3H2O) formed during the aromatization reaction
was measured by counting the radioactivity in the supernatant.
Unfixed frozen sections of the breast tissue excised from the propositus during reductive mammoplasty were cut into 5-µm thick sections and mounted on slides coated with chrome-alum gel, which were processed and covered with monoclonal mouse antiaromatase antibody prepared against human placental aromatase (provided by E. Simpson, University of Texas Southwestern Medical Center, Dallas, TX), as recently described (39).
DNA, mRNA, and protein studies
Genomic DNA was prepared from the patients and normal controls, as previously described (36, 40, 41). PCR amplification of the P450arom gene was performed with intronic primers previously reported (15, 16, 42, 43) or synthesized from the known untranslated sequences upstream of the P450arom gene (13, 15, 16, 19, 21). The products of these reactions were run on ethidium bromide-stained 1% agarose gels and were found to have the expected size, excluding major gene deletions or rearrangements (data not shown). Single strand conformation polymorphism analysis was performed as previously described (41). Direct and asymmetric PCR sequencing was performed as we previously reported (40, 41), using either the fmol sequencing method (Promega Corp., Madison, MI) or the Sequenase version 2.0 kit (U.S. Biochemical Corp., Cleveland, OH).
cDNA synthesis was accomplished from total RNA extracted from fibroblasts, PBL, EBV-tL of the patients and two normal control cell lines, and the breast tissue of the propositus (methods described in Refs. 40 and 41), with the antisense primer 24 (15) and the RT-PCR kit (Boehringer Mannheim, Mannheim, Germany). The sense primers 1a, 1b, 1c, and 1d and the antisense primer 2d (sequences listed in Ref.16) were used for the subsequent PCR amplification of the cDNA synthesized from cells from patients and normal controls and from the breast tissue of the propositus. The conditions of the reactions have been previously reported (16).
Rapid amplification of cDNA ends (RACE) was performed after mRNA extraction from EBV-tL (Clontech Laboratories, Palo Alto, CA) using the Marathon cDNA amplification kit (Clontech Laboratories). The sequences of the primers used for this reaction have been published (15, 16). Identification of the cDNA products was performed by cloning into the PCR-II vector, using the TA cloning kit (Invitrogen, San Diego, CA), and subsequent plasmid sequencing, or after a second round of PCR amplification and asymmetric PCR sequencing, using nested primers, as previously described (40, 41). Several PCR products and plasmids were also sequenced with the use of the ABI PRISM dye terminator reaction kit (Perkin-Elmer, Norwalk, CT) in an ABI PRISM 377 automatic DNA sequencer (Perkin-Elmer).
Northern blot analysis was performed after the identification of a
novel promoter in affected subjects. The multiple tissue Northern blot
(Clontech Laboratories) kit was used for this purpose. A
25-nucleotide-long DNA probe was synthesized and end labeled with the
use of T4 polynucleotide kinase (New England Biolabs, Beverly, MA) and
[
-32P]deoxy-ATP, as previously described (36).
Western blot analysis was performed with protein extracted from EBV-tL derived from the patients and normal controls, as previously described (40). Protein lysates were resolved by 8% SDS-PAGE, transferred to nitrocellulose membrane (Novex, San Diego, CA) and incubated with rabbit polyclonal antiserum, which was raised against human placental aromatase, purified, and characterized by Bellino et al. (44). This antibody recognizes a single band of 53,000 kDa, the approximate molecular mass of P450arom.
Genetic linkage analysis was performed with a polymorphic marker
(tetranucleotide repeat) located within the human P450arom gene on
chromosome 15q21.1 (45). Genomic DNA from the propositus, his sister,
his parents, and his paternal grandparents was amplified by PCR
(conditions of the reaction and primer sequences given in Ref.45), in
which one of the primers was end labeled with
[-32P]deoxy-ATP, as previously described (36). The
reactions were run on an acrylamide denaturing gel, which was then
subjected to autoradiography. Linkage analysis was performed with the
LINKAGE (version 5.1) package of computer programs (36, 46), assuming a
dominant model of inheritance, 100% penetrance in both sexes, and the
reported allele frequencies (45).
Ribonuclease protection assay (RPA) was performed with the use of a kit (RPA II, Ambion, Austin, TX), a plasmid constructed in our laboratory following the amplification of normal human genomic DNA with the 1c and 2d primers of the P450arom gene (sequences given in Ref.16), unidirectional subcloning into a TA vector (Invitrogen), and total RNA extracted from the proband’s EBV-tL cell line and two normal control EBV-tL cell lines. The protected fragments were run in a 6% acrylamide gel and autoradiographed.
Results
Clinical evaluation and treatment
Imaging studies were negative for any tumor in either child.
Pelvic ultrasound of the propositus’ sister revealed prepubertal
ovaries (<0.5 mL bilaterally), but endometrial thickening compatible
with estrogen stimulation. Table 1
summarizes the remaining findings. In both children, skeletal age was
advanced. Accordingly, the GnRH stimulation test results showed that
both children were in central precocious puberty. They also had high
serum E2 and, in particular,
E1 levels. Their father had high
E2 and E1 levels; his T and
dihydrotestosterone levels were in the low normal range at baseline and
responded normally to stimulation with hCG. Baseline
dehydroepiandrosterone sulfate levels were low for pubertal stage in
both children. The high baseline E2 and
E1 levels did not increase 1 h after ACTH
stimulation; hCG stimulation, on the other hand, produced significant
increases in both hormones in the propositus and his father. Their
response pattern, however, differed from that previously reported for
normal males (29, 30, 48), in that there was no immediate increase in
the two hormones (data not shown). Thus, E2 and
E1 increased after the first 24 h, after a
brisk elevation of A and T levels. These data and the relatively lower
baseline A and dehydroepiandrosterone sulfate levels were consistent
with extraglandular aromatization of C19 to C18
steroids in the propositus, his sister, and their father and excluded
other abnormalities of adrenal or gonadal steroidogenesis. Because the
children’s final height prediction was suboptimal (Table 2), they were both started on treatment
with a GnRH analog (histrelin acetate; 10–15 µg/kg, sc, daily), and
6 months later an aromatase inhibitor (testolactone; 40 mg/kg·day,
orally) was added, according to similar regimens used in children with
McCune-Albright syndrome and other disorders associated with precocious
puberty (34, 35). This treatment was successful in suppressing central
precocious puberty and arresting skeletal age as well as decreasing
serum estrogen levels in both children, as the most recent (3 yr into
the treatment) visit showed (Table 2). Gynecomastia in the propositus
was treated by bilateral reductive mammoplasty.
|
|
No aromatase activity was detected in nontransformed PBL from the
patients and normal controls, although aromatase gene transcripts were
detectable (see below) (50a). Increased enzyme activity was, however,
found in the cultured skin fibroblasts and EBV-tL derived from
clinically affected members of the family. The results are shown in
Table 3 and Fig. 1E
.
|
, upper
panel, a and c). Surrounding adipose fibroblasts were also
positively stained, but not as strongly. When the antiaromatase
antibody was replaced by an irrelevant mouse IgG, no immunoreactivity
was seen in the same tissue (Fig. 2, upper panel, b and
d).
|
, lower panel). Linkage analysis
To investigate whether the responsible genetic defect in this
kindred was located in the P450arom gene, a polymorphic,
tetranucleotide [(TTTA)n] repeat marker at nucleotide
position 682 of this gene (12, 13, 45) was used. Indeed, the marker
cosegregated with aromatase excess in affected members of this family
(Fig. 1D). With a model of autosomal dominant inheritance, the allele
frequencies previously reported (45), and a 100% penetrance for the
biochemical expression of increased aromatase activity, the combined
logarithm of odds score was 0.6, with no observed recombinations
[recombination fraction (
) = 0.00]. This represented a chance of
approximately 4.6 to 1 that the genetic defect in this family was
indeed located in the P450arom gene locus on chromosome 15q21.1.
Sequence analysis of the 5'-end of the P450arom gene
As the genetic defect in this family was likely to be in the P450arom gene, we tested the hypothesis that the former might represent a sequence change in the known regulatory regions of this gene’s promoter region, in particular those that have been found to have a silencing function on the expression of the gene (19). Single strand conformation polymorphism analysis of the known genomic sequences of the P450arom gene (15, 16, 19, 42) in the patients showed no pattern alterations that segregated with the disease, or that were not present in normal control DNA specimens (data not shown). Subsequent sequencing of the genomic DNA surrounding the placental (exon 1a or PI.I) and ovarian (exon 1d or PII) promoters (sequences reported in Refs. 13 and 19, respectively) revealed changes that did not segregate with the disease or were present in normal control DNA samples. The most extensive alteration was that at nucleotide 482 of the ovarian promoter region. The reported sequence at this location (19) reads (the numbers refer to cDNA position, the sequences altered in comparison to the published version are underlined). (478)5'-TTGATCCAATG-3'(488), whereas the following sequence was found in all members of our family and in 20 normal control DNA samples (478)5'-TTGCCTAAATG-3'(488) (data not shown). Similarly, a nucleotide substitution was present 15 bp upstream from the TATA box of the placental promoter in all our patients and in 20 control DNA specimens; the reported sequence (13) was [-45]5'-GACAATAAATG-3'[-35], and the sequence found was [-45]5'-GACAACAAATG-3'[-35] identical to that reported in Ref. 16 (data not shown).
These data showed that there were no disease-causing alterations in the studied promoter regions of the P450arom gene. The latter, however, cover a genomic area that is approximately 10,000 bp long, and the DNA clones containing the various promoters do not, currently, overlap; thus, a gap of unknown size is present in the genomic sequence of the 5'-end of the P450arom gene, which could harbor the defect in this family (11, 16).
Studies of P450arom mRNA: identification of a novel exon 1
To define the type of P450arom mRNA expressed in the cell lines of
our patients, we performed RT-PCR analysis with known primers of the
various promoters for this
gene,2 which showed that the
proximal promoters c and d (14) or PI.3 and PII (11) (Fig. 3A) were present in P450arom cDNA cloned
from PBL, EBV-tL, and cultured skin fibroblasts from the patients and
breast tissue from the propositus (data not shown). However, the
corresponding normal control cells also contained the same cDNA
species. The relative amount of P450arom cDNA present in PBL and EBV-tL
cell lines from normal controls was lower than that in our patients
(data not shown), consistent with their lower aromatase activity, as
mentioned above. Nevertheless, these data confirmed earlier evidence
that the P450arom gene was weakly expressed in peripheral leukocytes of
normal individuals (51) (Vottero, A., L. Kirschner, and C. Stratakis,
unpublished data).
|
). This novel exon 1 of the P450arom cDNA was
present in transcript pools of both PBL and EBV-tL from the affected
members of our kindred, but not in those established from normal
control PBL or EBV-tL; instead, the d (PII promoter) was cloned from
the latter (Fig. 3B). The sequence of the novel exon 1 is presented in
Fig. 3C and compared to that of exon 1d. This novel transcript was also
present in the fibroblasts of the affected members of the family and
the breast tissue of the propositus (data not shown). When a DNA probe
was made with the novel sequence, it hybridized with normal control
human polyadenylated RNA primarily from the ovary and secondarily from
the testis and small intestine, identifying a band of the expected
P450arom mRNA size (
3.4 kilobases; Fig. 3D). The RPA identified the
new cloned product in mRNA from the cell line of the propositus and
excluded the possibility that this was an artifact of the RACE
procedure (Fig. 3E); overexpression of the P450arom gene was confirmed
by the same assay in this cell line vs. control transfer RNA
(data not shown). Sequencing of the genomic DNA around the splicing
junction of the 5'-end of exon II of the P450arom gene in our patients
did not show any mutations that could be responsible for altered
splicing on this side (data not shown). However, as the entire sequence
of the 5' end of the P450arom gene is unknown, it is possible that the
genetic defect(s) affecting P450arom mRNA splicing in this family is
located within the gap that currently separates the known promoter
regions of this gene (Fig. 3A). Discussion
In the present study, a boy with prepubertal gynecomastia and his sister with isosexual precocious puberty had elevated serum estrogen levels. The source appeared to be nongonadal conversion of adrenal androgens (A and T) to estrogens (E1 and E2, respectively). Indeed, symptoms were manifested in both children around the time of or shortly after adrenarche. Although increased extraglandular aromatization of C19 steroids has been reported as a cause of gynecomastia twice before (7, 8), this is the first time that the aromatase excess syndrome has been associated with clinical manifestations in females. The first patient, reported by Hemsell et al. (7), was a 10 7/12-yr-old boy, who presented with a 2-yr history of gynecomastia and was in puberty at the time of his evaluation. Berkovitz et al. (8) subsequently reported five affected males, all members of the same family, who developed gynecomastia at an early age (10–11 yr). Because no females were affected in this kindred, and the disorder was present in a maternal uncle, it was speculated, as in yet another family (52), that aromatase excess was inherited in an X-linked manner, although the P450arom gene is on an autosomal and not a sex chromosome (chromosome 15q). Thus, autosomal dominant inheritance with sex-limited expression could not be excluded in these pedigrees. In our kindred, the aromatase excess syndrome was transmitted to the two children from their father, who had peripubertal gynecomastia, elevated serum estrogen levels as an adult, and increased aromatase activity in vitro. The male to male transmission and expression of the disease in a female are compatible with autosomal dominant inheritance in this family, suggesting that the aromatase excess syndrome may be a genetically heterogeneous disorder.
Despite increased estrogen levels, males with aromatase excess appear to be fertile (6). Indeed, the father of the propositus in our family was fertile and had normal libido despite a small testicular volume (15 mL bilaterally), and a testosterone level that was in the lower end of the normal range. Conversely, a male patient with aromatase deficiency had a large testicular volume (>25 mL bilaterally) and excessive testosterone levels, probably due to excess gonadotropins (53). In the aromatase excess syndrome, mild suppression of testicular growth and Leydig cell function probably reflect a direct estrogen negative feedback effect on pituitary gonadotropins (54, 55, 56). In general, the inhibitory effects of estrogen on reproductive function appear to be milder in males with this syndrome than in patients receiving exogenous estrogens or with estrogen-secreting tumors, probably because serum E2 and/or E1 levels in the former are lower than those in the latter.
Increased aromatase activity in females appears to be associated with isosexual precocity at a young age and, very likely, varying degrees of macromastia in adulthood. These findings are consistent with the major functions of estrogens, the growth and maturation of female secondary sexual characteristics, including the breast (56). Sexual function, fertility, and menstrual periods were normal in the paternal grandmother of our patients, who had increased aromatase activity in vitro. These findings should be contrasted with the dramatic effects of aromatase deficiency in female patients, which include virilization at birth, followed by hypergonadotropic hypogonadism, and sexual infantilism at puberty (53, 56, 57, 58, 59).
Both the boy and the girl with the aromatase excess syndrome had greatly advanced bone age; their degree of skeletal maturity was disproportionate to their height (chronological age < height age < skeletal age) despite accelerated growth. This situation is seen frequently in untreated precocious puberty, when the rate of skeletal maturation overcomes linear growth (56). The short stature of their father was also indicative of the potent effect on estrogens on accelerating premature epiphyseal closure. These findings are in sharp contrast with the tall stature and delayed epiphyseal closure of patients with aromatase deficiency (53, 57, 58, 59) or estrogen resistance (60, 61). Because of their predicted final height, our patients were treated with an aromatase inhibitor and a GnRH analog; although the 3-yr evaluation indicated an improvement in their final height prediction, the final outcome of this therapy remains to be seen.
In our kindred, the disorder cosegregated with a polymorphism of the P450arom gene, indicating that it was likely that the responsible genetic defect was located in this gene. Aromatase assays in fibroblast and EBV-tL cell lines established from the patients were in accordance with the clinical and biochemical data of aromatase excess, although they did not indicate whether other tissues were involved in the excessive aromatization. Aromatase activity on EBV-tL appeared to be a useful diagnostic tool in our patients; although P450arom mRNA was present in their peripheral leukocytes, aromatization in vitro was measurable only after transformation of the latter. Although the in vivo source of estrogen could be determined only with an infusion study, the hCG and ACTH stimulation tests and the 3-yr follow-up evaluation of our patients suggested that most of their aromatization took place in extragonadal tissues, which included the breast and, perhaps, the adrenal glands.
The P450arom protein in these cell lines had the expected mol wt, but was present in increased amounts compared to that in controls. It was likely that the situation in EBV-tL mirrored the pathogenesis of the disorder; an abnormality that led to overexpression of the gene in the tissues of our patients was probably further unmasked in their transformed lymphocytes.
Interestingly, the breast of our patient showed strong P450arom immunostaining in the glandular epithelium; adipose fibroblasts were stained, albeit less strongly, in the surrounding stroma. This pattern is the reverse of that recently observed in a female patient with macromastia and evidence of aromatase excess (62), but is consistent with other findings in cases of male patients with gynecomastia, where little or no stroma immunostaining was observed (63). We speculate that the distribution of P450arom expression in the breast may be gender, pubertal stage, or fat mass dependent. Males, for example, may have a higher expression in the glandular epithelium than in the surrounding stroma. Our patient was a very thin male, early in puberty (TS-II by testicular size), with a firm TS-IV breast. It is possible that the observed pattern of expression would have been different if there was lipomastia or if the breast was examined at a later stage. Finally, the different pattern of expression of the P450arom gene in our patient may just be due to the disease process, driven by the genetic defect responsible for the disorder in his family.
Due to its complex alternative splicing, the 5'-end of the P450arom gene has been considered an obvious candidate for disorders with increased aromatase activity (11, 16, 23, 24). A preliminary screen of the coding region of the gene and its two known promoter regions did not show any alterations. However, after a search for novel 5'-untranslated ends of the P450 aromatase gene by the RACE method, a novel exon 1 of the P450arom cDNA was cloned from the cell lines that demonstrated aromatase excess. This sequence appeared to be expressed only in gonadal tissues and the small intestine normally. In patients with the aromatase excess syndrome, however, it was present in their cultured skin fibroblasts, lymphocytes, and breast tissue. The cloned sequence is located in the genomic sequence neighboring the 5'-end of exon 2 of the P450arom gene and is a novel product of alternative splicing occurring in this area of the P450arom transcript. It is possible that pathological expression is present in other tissues as well, in particular in the adrenal glands, as symptoms of hyperestrogenemia in our patients appeared at the time of adrenarche. This is also supported by the modest effect that complete gonadal suppression had on E1 levels in these patients over a period of 3 yr.
It is likely that the genetic defect in the aromatase excess syndrome
is a sequence change in the P450arom promoter region. The latter either
causes a splicing error or affects the binding and/or activity of a
transcription factor that regulates P450arom mRNA synthesis. Under both
of these conditions, the tissue-specific selection of the various
promoters available for this gene could be altered in favor of the
newly identified cDNA, leading to inappropriate expression of the gene.
Analogous is the situation in the disorder known as hereditary
persistence of fetal hemoglobin, in which nucleotide substitutions in
the region immediately upstream from the transcription start site
affect the binding of trans-acting proteins involved in the
normal developmental repression of -hemoglobin gene expression (64).
Alternatively, a mutation in a regulatory factor of the P450arom gene
may be responsible for increased expression of the P450arom gene in our
patients. Under both hypotheses, excess aromatization would take place
not just in the tissues studied (skin fibroblasts, lymphocytes, and
breast), but also, and perhaps predominantly, in other tissues,
including the adrenal glands and, to a lesser degree, the gonads.
We conclude that the familial aromatase excess syndrome is a genetically heterogeneous disorder, which can be inherited in an autosomal dominant manner and lead not only to heterosexual precocity and gynecomastia in males, but also to isosexual precocity and macromastia in females. Aromatase expression can be shown in skin fibroblasts and EBV-transformed lymphocytes from affected patients in vitro and appears to be associated with the utilization of a novel exon 1 of the P450arom cDNA. The genetic defect responsible for this condition is likely to be located in the 5'-end of the P450arom gene.
Footnotes
1 This work was supported in part by NIH Grant CA-62483 (to
A.B.). 
2 The nomenclature of the various aromatase first
exons and corresponding promoters is unsettled. In this manuscript, we
use the codes established by Harada et al. (16 ); exons
1a, 1b, Ic, and Id correspond to promoters P1.1 (expressed in
placenta), P1.4 (adipose tissue), PI.3 (adipose tissue), and PII
(gonads) described by Simpson et al. (11 ),
respectively.
Received December 12, 1996.
Revised September 15, 1997.
Revised December 3, 1997.
Accepted December 9, 1997.
References
-hydroxyprogesterone,
4-andonestedione, estrone, and 17ß-estradiol: evidence
for the action of human chorionic gonadotropin on intermediate enzymes
implicated in steroid biosynthesis. J Clin Endocrinol Metab. 49:284–291.[Medline]
-globin gene in Greek hereditary persistence of fetal
hemoglobin. Nature 313:323–325.
This article has been cited by other articles:
 |
|
 |
|
  I. Czajka-Oraniec, W. Zgliczynski, A. Kurylowicz, M. Mikula, and J. Ostrowski Association between gynecomastia and aromatase (CYP19) polymorphisms Eur. J. Endocrinol., May 1, 2008; 158(5): 721 - 727. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. I. Shulman, G. L. Francis, M. R. Palmert, E. A. Eugster, and for the Lawson Wilkins Pediatric Endocrine Society Use of Aromatase Inhibitors in Children and Adolescents With Disorders of Growth and Adolescent Development Pediatrics, April 1, 2008; 121(4): e975 - e983. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Demura, R. M. Martin, M. Shozu, S. Sebastian, K. Takayama, W.-T. Hsu, R. A. Schultz, K. Neely, M. Bryant, B. B. Mendonca, et al. Regional rearrangements in chromosome 15q21 cause formation of cryptic promoters for the CYP19 (aromatase) gene Hum. Mol. Genet., November 1, 2007; 16(21): 2529 - 2541. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A Vottero, V Rochira, M Capelletti, I Viani, L Zirilli, T M Neri, C Carani, S Bernasconi, and L Ghizzoni Aromatase is differentially expressed in peripheral blood leukocytes from children, and adult female and male subjects. Eur. J. Endocrinol., March 1, 2006; 154(3): 425 - 431. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. E. Bulun, Z. Lin, G. Imir, S. Amin, M. Demura, B. Yilmaz, R. Martin, H. Utsunomiya, S. Thung, B. Gurates, et al. Regulation of Aromatase Expression in Estrogen-Responsive Breast and Uterine Disease: From Bench to Treatment Pharmacol. Rev., September 1, 2005; 57(3): 359 - 383. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Tiulpakov, N. Kalintchenko, T. Semitcheva, A. Polyakov, I. Dedov, P. Sverdlova, G. Kolesnikova, V. Peterkova, and P. Rubtsov A Potential Rearrangement between CYP19 and TRPM7 Genes on Chromosome 15q21.2 as a Cause of Aromatase Excess Syndrome J. Clin. Endocrinol. Metab., July 1, 2005; 90(7): 4184 - 4190. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. Binder, D. I. Iliev, A. Dufke, M. Wabitsch, R. Schweizer, M. B. Ranke, and M. Schmidt Dominant Transmission of Prepubertal Gynecomastia Due to Serum Estrone Excess: Hormonal, Biochemical, and Genetic Analysis in a Large Kindred J. Clin. Endocrinol. Metab., January 1, 2005; 90(1): 484 - 492. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Gennari, R. Nuti, and J. P. Bilezikian Aromatase Activity and Bone Homeostasis in Men J. Clin. Endocrinol. Metab., December 1, 2004; 89(12): 5898 - 5907. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. M. Martin, C. J. Lin, M. Y. Nishi, A. E. C. Billerbeck, A. C. Latronico, D. W. Russell, and B. B. Mendonca Familial Hyperestrogenism in Both Sexes: Clinical, Hormonal, and Molecular Studies of Two Siblings J. Clin. Endocrinol. Metab., July 1, 2003; 88(7): 3027 - 3034. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Shozu, S. Sebastian, K. Takayama, W.-T. Hsu, R. A. Schultz, K. Neely, M. Bryant, and S. E. Bulun Estrogen Excess Associated with Novel Gain-of-Function Mutations Affecting the Aromatase Gene N. Engl. J. Med., May 8, 2003; 348(19): 1855 - 1865. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. N. Kalantaridou and G. P. Chrousos Monogenic Disorders of Puberty J. Clin. Endocrinol. Metab., June 1, 2002; 87(6): 2481 - 2494. [Full Text] [PDF]
|
|
|
|
|
|
P. G. McTernan, L. A. Anderson, A. J. Anwar, M. C. Eggo, J. Crocker, A. H. Barnett, P. M. Stewart, and S. Kumar Glucocorticoid Regulation of P450 Aromatase Activity in Human Adipose Tissue: Gender and Site Differences J. Clin. Endocrinol. Metab., March 1, 2002; 87(3): 1327 - 1336. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. A. Stratakis, L. S. Kirschner, and J. A. Carney Clinical and Molecular Features of the Carney Complex: Diagnostic Criteria and Recommendations for Patient Evaluation J. Clin. Endocrinol. Metab., September 1, 2001; 86(9): 4041 - 4046. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. Van den Berghe, F. Weekers, R. C. Baxter, P. Wouters, A. Iranmanesh, R. Bouillon, and J. D. Veldhuis Five-Day Pulsatile Gonadotropin-Releasing Hormone Administration Unveils Combined Hypothalamic-Pituitary-Gonadal Defects Underlying Profound Hypoandrogenism in Men with Prolonged Critical Illness J. Clin. Endocrinol. Metab., July 1, 2001; 86(7): 3217 - 3226. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
