| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Prince Henry’s Institute of Medical Research (L.O., K.M.R., M.E.J., E.R.S.) and Department of Biochemistry (K.M.R.), Monash University, Clayton, 3168, Victoria, Australia
 |
Abstract |
|---|
 Top Abstract I. IntroductionII. Overview of SpermatogenesisIII. Biosynthesis and Action...IV. ERs, Aromatase, and...V. The Effects of...VI. Estrogen and Spermatogenesis...VII. Estrogen and Sexual...VIII. SummaryReferences |
|---|
and/or ß and aromatase are also compared to
evaluate the physiological role of estrogen in male fertility. The
review focuses on the effects of estrogen administration or
deprivation, primarily in rodents, on the hypothalamo-pituitary-testis
axis, testicular function (including Leydig cell, Sertoli cell, and
germ cell development and function), and in the development and
function of the efferent ductules and epididymis. The requirement for
estrogen in normal male sexual behavior is also reviewed, along with
the somewhat limited data on the fertility of men who lack either the
capacity to produce or respond to estrogen. This review highlights the
ability of exogenous estrogen exposure to perturb spermatogenesis and
male fertility, as well as the emerging physiological
role of estrogens in male fertility, suggesting that, in this local
context, estrogenic substances should also be considered "male
hormones." I. Introduction
II. Overview of Spermatogenesis
A. Germ cell development
B. Regulation
III. Biosynthesis and Action of Estrogen
A. Estrogen biosynthesis
B. Mechanisms of estrogen action
IV. ERs, Aromatase, and Estrogen Production in the Testis
A. ERs and aromatase in the fetal testis
B. ERs and aromatase in the immature testis
C. ERs and aromatase in the adult testis
D. Estrogen production in the testis
V. The Effects of Estrogen Administration or Deprivation on Spermatogenesis
A. Hypothalamo-pituitary-testis axis
B. Efferent ductules and epididymis
C. Testicular descent
D. Leydig cells
E. Sertoli cells
F. Germ cells
G. Comparison of the spermatogenic phenotype of mice with targeted disruptions of ERs or aromatase
VI. Estrogen and Spermatogenesis in Humans
VII. Estrogen and Sexual Behavior
VIII. Summary
|
I. Introduction |
|---|
|
TopAbstractI. IntroductionII. Overview of SpermatogenesisIII. Biosynthesis and Action...IV. ERs, Aromatase, and...V. The Effects of...VI. Estrogen and Spermatogenesis...VII. Estrogen and Sexual...VIII. SummaryReferences |
|---|
(ER)
(3, 4, 5, 6), or aromatase (7, 8, 9), along with the discovery of a second
estrogen receptor-ß (ERß) (10, 11, 12, 13), which is widely expressed in
the male reproductive tract (14). Another important reason for the increased interest in the role of estrogen in male reproduction stems from various reports that exposure to estrogens in the environment may have a detrimental effect on male reproductive development and health and may be related to the reported decreases in sperm counts over the past 50 yr (15, 16, 17). Although controversy exists as to whether male fertility over the past five decades has truly declined (18), and whether the exposure of humans to the relatively low levels of estrogens in the environment would cause significant health problems (19), research into the effects of estrogens on the male reproductive system is clearly warranted (19, 20). In agreement with the notion that estrogen exposure during development may impair male fertility, it has been known for many years that estrogen administration to experimental animals during the neonatal period or adulthood can impair sperm production and maturation (21, 22).
Normal male fertility relies on normal spermatogenesis, i.e., the process by which immature spermatogonia within the testis divide and differentiate into the mature elongated spermatid form that is subsequently released from the seminiferous epithelium. The full fertilizing potential of the released spermatozoan is also dependent on the progression and maturation of sperm through the excurrent duct system and the epididymis.
While the administration of estrogens and xenoestrogens during fetal
and neonatal development has been reported to be associated with a
series of male reproductive disturbances, such as cryptorchidism,
epididymal defects, impaired fertility, and an increased incidence of
testicular cancer (e.g., Refs. 23, 24, 25 and references
therein), an essential physiological role for estrogen in male
fertility was not identified until the early 1990s. The infertility in
mice lacking a functional ER (ERKO) was the first definitive
demonstration that estrogen was required for male fertility (3, 4, 5, 6).
ERKO mice are infertile primarily due to a defect in efferent
ductule development and function (6, 26). Mice lacking a functional
aromatase gene (aromatase knockout, ArKO) are also infertile; however,
this appears to be primarily due to a specific defect in germ cell
development (8). Thus lessons provided by mice transgenic for ER
and/or ERß [i.e., ERKO, ERßKO, and ERßKO mice
(see Ref. 27 for review)] as well as aromatase [ArKO mice (8, 9)],
provide compelling evidence for a role for estrogen in spermatogenesis
and male fertility.
Estrogen is clearly involved in numerous processes in the male, including bone turnover, behavior, and the cardiovascular system (see Ref. 2 for review). However, this review will focus on the effects of estrogen administration and deprivation on the production and function of spermatozoa. This review will address two important questions: 1) can exposure to estrogen or estrogenic substances interfere with spermatogenesis and male fertility? and 2) does estrogen have a physiological role in spermatogenesis and male fertility? Given that spermatogenesis relies on, among other things, the normal development and function of the hypothalamo-pituitary-testis axis, testicular cells (including Leydig cells, Sertoli cells, and germ cells), efferent ductules, and epididymis, the experimental evidence for the effects of, and/or a role for, estrogen in each of these systems will be presented. This review will also provide an overview of estrogen biosynthesis and action in general, and specifically in the testis. A brief overview of the phenotype of humans with mutations in ERs or aromatase will also be presented, as will an overview of a role for estrogen in sexual behavior, given that this is proving to be an important component of estrogen action in male fertility.
|
II. Overview of Spermatogenesis |
|---|
|
TopAbstractI. IntroductionII. Overview of SpermatogenesisIII. Biosynthesis and Action...IV. ERs, Aromatase, and...V. The Effects of...VI. Estrogen and Spermatogenesis...VII. Estrogen and Sexual...VIII. SummaryReferences |
|---|
). When germ cell development is
complete, the mature spermatids are released from the Sertoli cells
into the tubule lumen, and proceed through the excurrent duct system,
known as the rete testis, until they enter the epididymis via the
efferent ducts (Fig. 1). During passage through the epididymis, the
spermatids undergo a series of biochemical changes to become the motile
spermatozoa capable of fertilization.
|
), is enclosed by a capsule called the
tunica. The interstitial tissue contains the blood and lymphatic
vessels, which are essential for the movement of hormones and nutrients
into, and out of, the testis. The most frequently encountered cell type
in the interstitium is the Leydig cell (28), which is primarily
involved in the secretion of androgens, notably testosterone, as well
as other steroids including estrogen.
Within the seminiferous tubules, the Sertoli cells reside on a basement
membrane, under which are the lymphatic endothelium and the peritubular
myoid cells (29). The structure of the Sertoli cell is extremely
complex, with numerous cup-shaped processes encompassing the various
germ cell types (30) (Fig. 1). Developing germ cells form intimate
associations with Sertoli cells, with multiple germ cell types in
contact with one Sertoli cell. The various generations of germ cells,
as described below, are not randomly distributed within the
seminiferous epithelium, but are arranged in strictly defined cellular
associations (31). It is the unique associations of these germ cells
with Sertoli cells that constitute the cycle of the seminiferous
epithelium, and each particular association of germ cells is referred
to as a stage. The number of stages of spermatogenesis in a particular
species is thus defined by the number of morphologically recognizable
germ cell associations within the testis; in the mouse there are 12
stages, in the rat there are 14, and in the human there are 6 (32).
A. Germ cell development
Germ cell development involves a complicated series of events, and
the various germ cell types can be distinguished on the basis of
morphology and the differential expression of proteins.
Spermatogonia are present between Sertoli cells, close to the basement membrane of the tubule. They are the most immature germ cells in the testis, and include type A spermatogonia, intermediate spermatogonia (found only in rodents), and type B spermatogonia, the latter of which are considered to be committed to differentiation. The true stem cells of the germ cell population are considered to be a subset of the type A spermatogonial population, although their identity cannot be discerned on the basis of morphology (33). Spermatogonia undergo numerous mitoses to produce a large number of germ cells available for entry into meiosis. Thus, proliferation of the spermatogonial population provides the source for the millions of sperm that are produced per day (32).
After the last mitosis of type B spermatogonia, preleptotene primary spermatocytes are formed (32). These cells replicate their DNA and hence initiate meiosis (33). During the prophase of the first meiotic division, germ cells undergo morphological transitions that can be classified on the basis of nuclear size and morphology (34). In the zygotene phase, pairing of homologous chromosomes occurs, and cells with completely paired chromosomes are termed pachytene spermatocytes. After the pachytene phase, a brief diplotene phase follows in which the chromosome pairs partially separate, and the cells then undergo the first meiotic division to yield secondary spermatocytes. These cells quickly undergo the second meiotic division to yield the haploid round spermatid.
The differentiation of round spermatids into the mature elongated spermatid form takes place, with no further division, during the process known as spermiogenesis. Briefly, spermiogenesis involves formation and development of the acrosome and flagellum, condensation of the chromatin, reshaping and elongation of the nucleus, and removal of the cytoplasm before release of the spermatid during spermiation (31, 32). After commencement of spermatid elongation, the highly condensed spermatid nucleus becomes incapable of transcription, and immature round spermatids transcribe high levels of mRNAs that are subject to translational delay until translation is required during elongation (35). Spermiation is the final step of spermiogenesis and involves the release of the mature elongated spermatid from the Sertoli cell into the lumen of the seminiferous tubule (36).
B. Regulation
Germ cell development relies on a highly coordinated interaction
with the Sertoli cell. Germ cells and Sertoli cells can communicate
directly via ligand/receptor-mediated interactions or paracrine
factors. The production and secretion of many Sertoli cell proteins
involved in germ cell development occur in a stage-dependent manner
(37), reflecting the ability of the Sertoli cell to adapt to the
changing needs of the germ cell. For many years it was presumed that
Sertoli cells were the major controlling factor in the timing of germ
cell development; however, recent studies investigating rat-to-mouse
spermatogonial transplantation clearly demonstrated that rat germ cells
in contact with mouse Sertoli cells develop according to the kinetics
of rat spermatogenesis, thus highlighting the role of germ cells
in controlling their own fate (38).
As well as the production of spermatozoa, the testis is involved in the production of hormones that are required for various functions in the body, including maintenance of secondary sexual functions, and feedback on the hypothalamus and the pituitary to control the secretion of the gonadotropins LH and FSH.
It is well known that the gonadotropins are the major endocrine regulators of spermatogenesis (39, 40, 41). LH targets the Leydig cell to stimulate the secretion of androgens, namely testosterone, which in turn acts on androgen receptors in the seminiferous epithelium to control spermatogenesis. FSH targets receptors within the Sertoli cell to regulate spermatogenesis by stimulating the production of numerous Sertoli cell factors. The roles of testosterone and FSH in the testis have been studied extensively, yet relatively little is known about how these hormones act within the Sertoli cell to stimulate and maintain spermatogenesis (39, 40, 41). Androgens alone have been shown to stimulate all phases of germ cell development in the hypogonadal (hpg) mouse, which is congenitally deficient in GnRH and therefore LH and FSH (42), highlighting the requirement of spermatogenesis for androgen. The question of whether FSH was essential for spermatogenesis in mice was answered by the generation of transgenic mice possessing targeted disruptions of the FSH receptor gene (43) or the FSH ß-subunit gene (44). Males of both transgenic models are fertile and display all stages of germ cell development, as are the androgen-treated hpg mice (42), suggesting that FSH is not an absolute requirement for fertility. However, in both cases the testes are smaller, and less sperm are produced (42, 43, 44), due to the requirement for FSH during the neonatal period of Sertoli cell division (45, 46). More recent quantitative studies on FSH receptor knockout mice also demonstrated defects in sperm development, leading to the production of poor quality sperm (47). Thus while FSH is not essential for spermatogenesis, it is clearly essential for quantitatively normal spermatogenesis and fertility. In terms of the endocrine regulation of spermatogenesis by FSH, LH, and androgens, it is clear that the initiation and maintenance of quantitatively normal spermatogenesis and thus full fertility rely on the delicate balance of the hypothalamo-pituitary-testis axis. The focus of this review is the growing body of evidence that suggests that estrogen should be added to the list of hormones involved in the regulation of spermatogenesis.
|
III. Biosynthesis and Action of Estrogen |
|---|
|
TopAbstractI. IntroductionII. Overview of SpermatogenesisIII. Biosynthesis and Action...IV. ERs, Aromatase, and...V. The Effects of...VI. Estrogen and Spermatogenesis...VII. Estrogen and Sexual...VIII. SummaryReferences |
|---|
The human CYP19 gene was cloned some years ago (50, 51, 52),
when it was shown that the coding region spans 9 exons beginning with
exon II. Upstream of exon II are a number of alternative exons I which
are spliced in the 5'-untranslated region of the transcript in a
tissue-specific fashion (Fig. 2). The
expression of the gene in the ovary and testis utilizes a proximal
promoter, promoter II (53, 54), and thus transcripts in gonadal tissues
contain sequence at their 5'-end that is immediately upstream of
the translational start site. In contrast, placental transcripts
contain at their 5'-end a distal exon, I.1, which is localized 100 kb
upstream from the start of translation in exon II. This is because
placental expression is driven by a powerful distal promoter upstream
of exon I.1 (55). Transcripts in adipose tissue contain yet another
distal exon located 20 kb downstream of exon I.1, exon I.4 (56), and a
number of other untranslated exons have been characterized including
one that is expressed in brain (57). Splicing of these untranslated
exons to form the mature transcript occurs at a common 3'-splice
junction that is upstream of the translational start site. This means
that although transcripts in different tissues have different
5'-termini, the coding region and thus the protein expressed in these
various tissue sites are always the same. However, the promoter regions
upstream of each of the several untranslated first exons have different
cohorts of response elements, and so regulation of aromatase expression
in each tissue that synthesizes estrogens is different. Aromatase
expression in the gonads is regulated by cAMP and gonadotropins due to
an interaction of the gonadal promoter II with the transcription
factors CREB (cAMP response element binding protein) and SP1 (58, 59).
For comparison, the adipose promoter, I.4, is regulated by class I
cytokines such as interleukin-6 (IL-6), interleukin-11 (IL-11), and
oncostatin M, as well as by tumor necrosis factor- (TNF). The
regulation of estrogen biosynthesis in each tissue site of expression
is unique and has been reviewed previously (60).
|
. The
two receptors are not isoforms of each other, but rather are the
products of distinct genes located on separate chromosomes.
The ER proteins are each composed of six functional domains labeled
A–F, a signature characteristic of the entire superfamily (Fig. 3). The N'-terminal A/B domain is the
least conserved among all members and demonstrates only 17% homology
between the two ERs. It contains the activation function 1 (AF1)
region, which is one of two regions critical for the transactivation
function of the members of the receptor family. By contrast, the C
domain is the most highly conserved region, being the DNA binding
domain that contains the zinc-finger motifs. The E domain, or
ligand-binding domain, is modestly conserved throughout the superfamily
and confers ligand specificity on the members. Conservation of amino
acid sequence in this region is 60% between the ER and ERß;
however, each binds estradiol with about equal affinity, although the
relative binding of other ligands differs substantially between them
(64, 65). The E domain also contains the major dimerization surface of
the receptors, and the second transactivation function, activation
function 2 (AF-2), is also located in this region of the
C'-terminus.
|
gene in vivo leads
predominantly to a single transcript of approximately 6.3 kb, encoding
a protein of 599 amino acids (66). The human ER is slightly shorter
at 595 amino acids (62). The existence of multiple promoter and
regulatory regions in the 5'-untranslated region has been reported, as
well as naturally occurring variants of the transcript in normal and
malignant tissues, but as yet the existence of true protein products of
these variants, as well as their significance, remains controversial
(67). Initial studies of the rodent ERß transcript indicated it was
substantially shorter than the ER, namely 485 amino acids (10, 11).
This is largely due to a significantly shorter N'-terminal region.
Indeed the existence of a functional AF-1 region in the ERß has been
questioned. However, open reading frames initiating upstream from that
first described have now been discovered (68), suggesting the
possibility that an ERß protein of approximately 530 amino acids
might also exist. Unlike the ER gene, Northern blots of rodent
tissues indicate the presence of multiple ERß transcripts, and a
number of variants have been described, including a conserved insertion
of 18 amino acids in a C'-terminal region in the rat, mouse, and human,
the deletion of one or more exons in these same species, and several
isoforms in the extreme C'-terminus of the human transcript (69, 70, 71).
There is considerable tissue specificity in the expression of ER and
ERß. Thus, ER is the dominant species expressed in uterus, liver,
adipose, skeletal muscle, pituitary, and hypothalamus, whereas ERß is
the major form in ovary and prostate, as well as other regions of the
brain including the limbic system, cerebellum, and cerebral cortex (72, 73). The localization of ER and ERß in the testis and epididymis
is discussed in Section IV.
The mechanisms of estrogen and estrogenic ligand action in the control of transcription have been recently reviewed elsewhere (74, 75, 76, 77, 78); however, the following will provide a brief overview of estrogen action on ER via classic pathways, ligand-independent ER activity, and nongenomic actions of estrogen.
The binding of ligand to ER and the events that lead to the regulation of target gene transcription are broadly similar among all steroid receptors, namely that they bind to response elements on the target gene and form a platform to which is recruited the complex of coactivators and transcription factors required for transcriptional activation. In general, estrogen action on ER involves ligand binding, dissociation of chaperone complexes and receptor phosphorylation; receptor dimerization; nuclear translocation; DNA binding and interaction with cofactors; and modulation of transcriptional activity.
Although ER and ERß bind estradiol with similar affinity, there is
considerable selectivity of the different receptor subtypes in terms of
affinities of various ligands (65). In particular, some phytoestrogens,
such as genistein and coumestrol, have a significantly higher affinity
for ERß than ER (64).
ER subtypes form homo- or heterodimers in vitro or in cells.
In general terms, the ER homodimer is more transcriptionally active
than the ERß homodimer in most systems (see Ref. 76 for review). The
ER/ERß heterodimer was recently shown to have similar
transcriptional activity to the ER homodimer under
hormone-saturating conditions, but is less active than the ER
homodimer at subsaturating estradiol concentrations (79).
The fact that there are distinct differences between the ER subtypes
with regard to agonistic/antagonistic effects on transcription is now
well known. For example, the ER agonistic activity of tamoxifen appears
unique to ER (see Ref. 76 for review). The distinct properties
exhibited by ER and ERß have been exploited for therapeutic use by
the development of selective ER modulators (SERMs) (see Refs. 74, 78
for review); however, an understanding of the molecular events
underlying SERM pharmacology has only recently begun to emerge (74, 78, 80). The molecular mechanisms underlying the differences in the
functional characteristics of ER and ERß are well beyond the scope
of this review, but include differences at the level of ligand
binding/affinity, cofactor recruitment, and activity of the AF-1 and
AF-2 domains (which are influenced by cell and promoter context) (see
Refs. 74, 75, 76 for review). For example, the relative expression levels
of the respective ER subtypes within a cell are key determinants of
transcriptional activity in response to agonists and antagonists (79).
Another way in which ER subtypes exhibit functional differences is in
their ability to recruit coactivators and corepressors (74, 75, 76, 80, 81). Examples of coactivators of ERs are the steroid receptor
coactivators (SRCs) (82) (see Ref. 83 for review). Often, coactivators
such as SRC1 facilitate transcription via histone acetylation activity,
which results in decondensation of chromatin and hence increases the
ability of the transcriptional machinery to interact with the promoter
(see Ref. 76 for review). In contrast to coactivators, corepressors
generally bind to ER in the absence of ligand or to antagonist-occupied
ER and repress transcriptional activity. These frequently possess, or
else activate, histone deacetylase activity. Examples of corepressors
include nuclear receptor corepressor (NCoR) (84) and short heterodimer
partner (SHP) (85, 86). The differences in transcriptional activation
of ER compared with ERß by agonists and antagonists is due, at
least in part, to differences in the ability of the ligand-receptor
complex to recruit coactivators, which is related to the conformational
changes induced in the helix H12 of the ligand binding cavity of each
receptor by the agonist/antagonist (80). In summary, the ability of an
ER subtype to stimulate transcription is dependent on promoter context,
the nature of the ligand bound, the expression of coactivator and
corepressor proteins, and the relative expression levels of ER subtypes
in a given cell. It is also worthwhile to note that the ability of a
ligand for the ER to influence transcription is determined by its
structure, its affinity for ER subtypes, and the recruitment of
coactivators and corepressors that is, in turn, dependent on cell and
promoter context (see Ref. 78 for review).
ERs can also associate with target gene promoters in a manner that is
not dependent on direct DNA binding, but involves the binding of ER to
proteins within a preformed transcriptional complex. For example,
ligand-activated ER can positively regulate transcription by
associating with the activating protein 1 (AP1) transcription factor
complex in target cells (see Ref. 77 for review). Similar interactions
of ER with other transcriptional complexes exist (see Ref. 74 for
review) indicating that ER can interact directly or indirectly with DNA
to modulate transcription. ER and ERß exhibit different properties
in both DNA binding-dependent and -independent mechanisms (see Ref.
76).
There is now also increasing evidence for ligand-independent
transcriptional activation of ERs (87, 88). Epidermal growth factor
(EGF)-induced phosphorylation of ER results in ligand-independent
transcriptional activity, a pathway that has been shown to have
functional significance (87, 89, 90). Growth factor-induced
phosphorylation of ER is dependent on the mitogen-activated protein
kinase (MAPK) pathway (see Ref. 91 for review). MAPK
phosphorylates the AF-1 region of ER (92) and ERß (93) and in the
latter case has been shown to promote the recruitment of SRC1 in a
ligand- independent manner (93). Other factors, such as cyclin D1,
can activate ER-mediated transcription in the absence of estradiol (94, 95).
In addition to the classic genomic pathway of ligand- occupied ER
interaction with target genes, estrogen can also induce extremely rapid
(within seconds to minutes) increases in the concentration of calcium
or cAMP second messengers (96, 97) in an apparently nongenomic
mechanism of action (see Ref. 98 for review), presumably via receptors
on the plasma membrane. Also, physiological concentrations of estradiol
can induce a rapid release of nitric oxide in endothelial cells via
membrane-bound receptors (99, 100). Nongenomic actions of estrogen
appear to be of particular importance in cardiovascular (101) and
neuronal (102) tissues. Evidence has been presented that in some
systems estrogen appears to act on a membrane-associated ER that is
immunologically related to ER (103, 104, 105, 106, 107). ERß transcripts can also
produce membrane-associated forms (104). However, in other systems,
nongenomic effects of estrogen do not appear to be via
membrane-associated forms of ER or ERß (108, 109, 110, 111). For example,
nongenomic actions of estrogen on pancreatic ß-cells appear to be via
a membrane binding site with the pharmacological profile of the
-adrenergic receptor (110). Therefore, the potential exists for
estrogen action on plasma membrane receptors/complexes unrelated to
ER or ERß.
In summary, ligand-bound ERs act as transcription factors and are
capable of modulating target gene transcription by both DNA
binding-dependent and -independent means. Molecular mechanisms of
estrogen action are now known to be more complex than originally
thought, with ligand-independent pathways that mediate cross-talk with
growth factor signaling pathways, as well as rapid nongenomic actions.
Understanding how estrogen will act within a cell is further
complicated by the fact that ER and ERß differ in their functional
characteristics and in their tissue localization. Thus, estrogenic
ligand-dependent and/or ER-dependent regulation of target gene
expression depends on the nature of the ligand, the relative levels of
ER subtypes and cofactors, the molecular pathway of action and, in
particular, the cell and promoter context.
|
IV. ERs, Aromatase, and Estrogen Production in the Testis |
|---|
|
TopAbstractI. IntroductionII. Overview of SpermatogenesisIII. Biosynthesis and Action...IV. ERs, Aromatase, and...V. The Effects of...VI. Estrogen and Spermatogenesis...VII. Estrogen and Sexual...VIII. SummaryReferences |
|---|
, ERß, and aromatase mRNA and protein in the
testis from the fetal stage through to adulthood will be presented. The
reader is also directed to other relevant reviews (113, 114, 115, 116) for the
discussion of ER and aromatase localization in the male reproductive
tract.
A. ERs and aromatase in the fetal testis
ERs and aromatase are found at all stages of testicular
development in the rodent. For more detailed information of male
gonadal differentiation, see Refs. 117, 118, 119 .
Immunohistochemical studies using antibodies that recognize ER show
that this protein is present in the mouse undifferentiated gonad at day
10 (120) and day 11.5 (121), suggesting that estrogen may have a role
very early in the differentiation process. Leydig cells within the
rodent fetal testis contain ER until birth (120, 121, 122, 123, 124, 125). In fact, ERs
are expressed in the Leydig cells at a stage in development when the
androgen receptor is not yet expressed (126), highlighting a role for
estrogen at this stage (see Ref. 125). Although one study reports that
ER is present within the seminiferous tubules of the fetal testis
(120), other studies do not (see Table 1). ER is also present in the
developing efferent ductules and epididymis (120, 121, 125).
|
), and this receptor subtype appears to remain in these cells until
birth (122, 123, 124, 127). However, it is the gonocytes that express ERß
in higher abundance than the other testicular cells (123, 127),
possibly suggesting a direct role in gonocyte maturation. ERß is also
present in rat associated ducts, which will eventually differentiate
into the efferent ductules and adjoining epididymis (124, 125). The fetal rat testis also has aromatase activity (128, 129, 130), which is first expressed by day 19 (129). The cell type responsible may be Sertoli cells, as estrogen production by fetal testes in culture is stimulated by FSH (131). Interestingly, this is close to the time when FSH receptors first appear on the Sertoli cells (132). There is also evidence that fetal Leydig cells have the capacity for LH-stimulated aromatization of androgens (133).
Thus the demonstration that rodent fetal testicular cells synthesize
estrogen and express both ER and -ß suggests a role for estrogen
in the development of the fetal male reproductive tract and in gonocyte
differentiation.
B. ERs and aromatase in the immature testis
Around the time of birth, the testis continues to express both ER
isoforms and aromatase (Table 1). Most reports suggest that ERß
protein and mRNA, rather than ER, appear to be localized to the rat
seminiferous epithelium, in both Sertoli cells and developing germ
cells (123, 127) (Table 1). In the mouse, the testis at this time also
continues to abundantly express ERß mRNA, with immunocytochemical
studies localizing this ER to the germ cells (122).
ER expression is restricted to the cells that lie outside of the
seminiferous epithelium shortly after birth (121, 125), with ER
(121, 122, 123, 124, 125), along with ERß (123, 127), continuing to be found in the
Leydig cells of mice and rats during the neonatal period. ER protein
is also found in the rodent rete testis (125), efferent ductules (121, 125), and the epididymis (121, 124, 125), where it has remained since
fetal development. In fact, ER is expressed in much higher levels in
the efferent ducts than the cauda epididymis (121, 125). ERß is
also found in the epididymis during development (122, 124, 134).
At this stage of development, basal aromatase activity is found in both the immature Leydig cells and Sertoli cells (135, 136, 137); however, activity is significantly induced by FSH in the Sertoli cells (135). In fact at this age, Sertoli cells were more active in producing estrogen than neonatal Leydig cells and adult Sertoli cells, suggesting that these cells are an important source of estrogen in the postnatal testis (137). At this age, germ cells have been reported not to contain detectable aromatase (138).
During days 10–26 in the immature rat, Leydig cells and Sertoli cells
are dividing and undergoing functional maturation (see Ref. 139 for
review). Again, ER is absent from the seminiferous epithelium with
only ERß prominent in the epithelium of the immature rodent (Table 1). Specifically in the tubules of the rat, ERß mRNA and protein are
present in relatively low levels in spermatogonia (123, 127) and
immature Sertoli cells (123, 127), and by day 21 ERß is abundantly
expressed in pachytene spermatocytes (123). Other germ cells do not
show ERß staining at this time (127). By day 12 in the mouse, ERß
is also specifically immunolocalized to the spermatocytes; however, its
expression appears to decrease and was undetectable by day 26 (122).
Again, rat and mouse Leydig cells express ER at this time (122, 123, 124, 125)
(Table 1).
During the neonatal and pubertal period of development, ER is
prominent in the rete testis and efferent ductules (125, 134) and is
present in the mouse epididymis (122). Recent studies on the
immunoexpression of ERs in the rat described the specific cellular and
regional localization of ER and ERß in the developing rat efferent
ductules, epididymis, and vas deferens and showed that the localization
of ER in the epididymis is confined to a relatively short window of
development (134). ERß is also present in the epididymis at this time
(122, 134).
Aromatase appears to have an age-dependent pattern of expression. As the animal matures, the Leydig cells appear to significantly supplement estrogen production from the Sertoli cells (136), with Leydig cell basal aromatase activity increasing 3- to 4-fold (137) and now stimulated by LH (135). However, the Sertoli cells do continue to express aromatase during their maturation (135).
C. ERs and aromatase in the adult testis
The presence of ER, ERß, and aromatase in the adult testis
has been the subject of numerous recent studies. An intense effort has
focused on the expression and localization of ER subtypes in the adult
testis of mice (73, 140), rats (14, 123, 125, 127, 141, 142), primates
(125, 143, 144), and humans (13, 145, 146, 147), primarily due to the fact
that there has been a considerable amount of conflicting data.
Various evidence indicates that Leydig cells of rats and mice express
ER (125, 142, 146) (see Table 1); however, the localization of ER
in Leydig cells of primates and humans is more controversial; studies
have shown immunostaining in some primate Leydig cells (125), and in
human Leydig cells (147); however, other studies could not detect ER
immunostaining in Leydig cells from primates (144, 146) or humans (146, 147). In the adult mouse ERß protein is expressed in the
Leydig cells (140), but this does not seem to be the case in the adult
rat (127, 142). Again, the localization of ERß in Leydig cells from
primates and humans is controversial with some studies showing no ERß
in Leydig cells from primates (143) or humans (13), whereas other
studies showed immunoexpression of ERß in Leydig cells from two
primate species (146) as well as humans (146, 147, 148).
By adulthood, rodent Leydig cells express a high level of aromatase (115, 138, 149, 150, 151), which is stimulated by LH (136, 152, 153) and steroids (153). In fact, aromatase activity is higher in the adult than at any other age (136) and is higher in the adult Leydig cells than in the Sertoli cells (149). The decrease in aromatase activity during Sertoli cell maturation into the adult form (see Ref. 154 for review) may be related to the control of Sertoli cell division and adult cell function (see Section V.E). The presence of aromatase in the Leydig cells of primates and humans is well established (see Ref. 115 for review).
Rat Sertoli cells contain both ERß mRNA and protein from the fetus to
adulthood (Table 1 and Section IV.A and B) (14, 123, 127, 141, 142). Immunohistochemical studies suggest that ERß is
not stage dependent in the rat (14, 123). In contrast to the rat, there
are no reports of ERß in mouse Sertoli cells (Table 1). Sertoli cells
in primates (146) and humans contain ERß (146, 147, 148) but not ER
(125, 146, 147).
There is now considerable evidence that germ cells contain both ERs and
aromatase (see Fig. 4 for a summary of
the proposed localization). In general, ERß is the predominant, and
potentially the only, ER in germ cells (see Table 1). One study in rats
found immunoexpression of ER in spermatocytes and spermatids (142);
however, there are no other reports of ER in rodent germ cells (see
Table 1). ER immunoexpression in human germ cells within
seminiferous tubule fragments in culture has been suggested by one
study (145), whereas other studies did not show ER protein in
primate or human germ cells (146, 147). ERß is present in
rat type A spermatogonia (127) as well as in intermediate and type B
spermatogonia in some studies (123), but not others (127) and is also
seen in spermatogonia in monkeys and humans (146, 148).
ERß is found in pachytene spermatocytes and round spermatids, but not
elongating spermatids in rats (123, 127, 141) and in primates and
humans (13, 143, 146, 148), although one study suggested ERß in human
elongating spermatids (145). However, one study in mice found ERß
only in elongated spermatids and not earlier germ cells (140) (see
Table 1). Other studies in rats and humans found ERß in Sertoli cells
but not in germ cells (142, 147). A further complexity to the
testicular localization of ERß is that one study in mice could not
detect ERß mRNA at all in mouse testis by RNase protection assay
(73). Therefore, considerable conflict in the literature exists in
terms of ERß localization in the testis (see Table 1 and below).
|
). Aromatase localization is
observed to move from the Golgi apparatus to the cytoplasm during
spermatid development (150). When round spermatids begin the
morphological transformation into elongated spermatids, aromatase
continues to be found in these cells (Table 1) and is immunolocalized
to the flagella of the developing spermatid (150). Aromatase appears to
be present in higher levels in mature spermatids of the rat than in
earlier germ cells (149, 155). Aromatase mRNA and activity was higher
in the germ cells of the mouse and rat when compared with Leydig cells
(150, 155, 157), suggesting that the germ cells are an important source
of estrogen in the testis.
When elongated spermatids are released from the epithelium, during the
process of spermiation, aromatase remains in the residual body that is
subsequently phagocytosed by the Sertoli cell (150, 157). However, not
all the cytoplasm is phagocytosed, and aromatase activity remains in
the cytoplasmic droplet that is still attached to the flagellum as the
sperm make their way through the epididymis (151, 158). Thus it appears
as if mature sperm are able to synthesize their own estrogen, as they
traverse the efferent ducts (116, 151, 158) (see Fig. 4). The ability
to synthesize estrogen gradually decreases as the droplet slowly moves
to the end of the tail during epididymal transit until it is finally
lost (158). The demonstration of aromatase in sperm is important as it
suggests that the sperm themselves could control the levels of estrogen
present in the luminal fluid, directly modulating functions such as the
reabsorption of fluid from the efferent ductules (116).
A very high level of expression of ER is seen in the efferent
ductules of the rat (125, 159). In fact, it has been found that it is
the efferent ductules that possess the highest level of ER
immunostaining, relative to the testis, excurrent ducts, and
epididymis, throughout life (125). In addition, the efferent ductules
appear to be the first male reproductive structure to express the ER in
fetal development (160), suggesting a role for estrogen in the
development of this tissue. ER is present in the rat and mouse
epididymis (73, 124, 159), although in low levels, while another report
could not detect it in rats (125). Again, similar to earlier ages, both
ERß mRNA and protein are present from the efferent ducts to the
epididymis of the mouse (73, 140) and the rat (124, 159). A very recent
study of the localization of ER and ERß in the rat efferent
ductules and epididymis showed that in adult rats ER was strongly
immunoexpressed in the epithelial cells of the efferent ductules but
was not found in the epididymis, whereas ERß was found in epithelial
cells and some periductal cells throughout the efferent ductules and
epididymis (134). ER is found in the nonciliated epithelial cells of
the efferent ductules in primates and humans (146), although another
study in humans did not observe ER immunoexpression in the efferent
ductules (147). ER was rarely detected in the epididymis of primates
or humans (146, 147). In contrast, ERß is found in both stromal and
epithelial cells throughout the efferent ductules and epididymis in
these species (146).
The studies reviewed in this section highlight the fact that there is
considerable conflict in the literature regarding the localization of
ER and ERß, particularly in the testis. The conflict usually
arises when comparing studies utilizing immunohistochemical techniques.
In the studies cited in Table 1 and above, there is considerable
variation in the antibodies used. The antibodies have been raised
against different regions of the ER molecules, and thus detect
different epitopes (see Ref. 161 for a recent review on ER and ERß
antibodies, including some of those used in the studies cited here).
Most studies cited here have used antigen retrieval methods and have
presented some evidence of antibody specificity, usually by
preabsorption controls. Even though various studies used commercially
available antibodies, it is clear that a range of commercial antibodies
to ERs are not well characterized (see Ref. 161 for review). In
addition, whether the antibody used detects ligand-occupied and/or
inactive receptor, monomeric or dimerized receptors, an ER in
heterodimer form, receptors bound to DNA, etc., is usually not known. A
further complexity in the immunohistochemical studies cited in this
review is that different methods of tissue fixation, antigen retrieval,
antibody dilution, and antibody detection methods were used, all of
which can affect the ability of an antibody to bind to its antigen.
Thus at present it is almost impossible to reconcile the differences in
the various reports on ER and ERß localization. Clarification of
the exact localization of ER subtypes in the testis would be
facilitated by the comparison of a panel of very well characterized
antibodies, in which the exact nature of the protein detected is known.
In summary, the studies reviewed above suggest that the testis is
capable of synthesizing and responding to estrogens throughout all
stages of development. The localization of ER, ERß, and aromatase
in the adult testis is summarized in Fig. 4 in an effort to bring
together the literature in this area. The localization of ER, ERß,
and aromatase demonstrates that estrogen action is likely to be
important for Leydig cell, Sertoli cell, and germ cell development and
function, as well as in the development and function of the efferent
ductules and epididymis. In particular, germ cells are capable of local
estrogen synthesis and response, via ERß, suggesting that paracrine
and intracrine actions of estrogens may be important in male germ cell
development. The localization of aromatase in sperm in the testis, and
as they traverse the efferent ductules and epididymis, together with
the demonstration of high levels of ER and ERß in the efferent
ductules, support the hypothesis that estrogen in sperm acts on ER in
the efferent ductules (see Section V.B).
D. Estrogen production in the testis
The synthesis of estrogens in the testis has been reviewed
extensively elsewhere (154, 162); therefore, only a few relevant
details will be considered here. There is a high concentration of
estrogen in rete testis fluid (163) and, in the rat, the concentration
of estrogen in the caput epididymis is approximately 25 times the level
measured in plasma (164). It is clear that the concentrations of
estrogen in the testis and rete testis fluid far exceeds the
concentration in male serum in various species (see Ref. 162 for
review) thus suggesting a central role for estrogen in testicular and
epididymal function.
In dissected testicular tissue from adult rats, the concentration of estrogen in interstitial tissue was 9 times higher than that in the seminiferous tubules (165). However, it is now becoming clear that the level of aromatase activity in germ cells of the adult rodent is equal to or higher than the aromatase activity in Leydig cells (149, 150) (see above), suggesting that while Leydig cells have previously been considered to be the primary source of estrogen in the testis (166), germ cells must now be considered to have an important role also (see Refs. 114, 115, 116 for review, and Section IV.C). Thus the source of the high concentration of estrogen in fluid leaving the testis may be due largely to the high levels of aromatase mRNA, protein, and activity in testicular germ cells and particularly in spermatids (149, 150, 151, 157, 158, 167, 168) (see Refs. 114, 115, 116 for review).
|
V. The Effects of Estrogen Administration or Deprivation on Spermatogenesis |
|---|
|
TopAbstractI. IntroductionII. Overview of SpermatogenesisIII. Biosynthesis and Action...IV. ERs, Aromatase, and...V. The Effects of...VI. Estrogen and Spermatogenesis...VII. Estrogen and Sexual...VIII. SummaryReferences |
|---|
A. Hypothalamo-pituitary-testis axis
As mentioned in Section II.B, the initiation and
maintenance of spermatogenesis require the secretion of gonadotropins
from the pituitary and thus is dependent on the balance of the
hypothalamo-pituitary-testis axis. The negative feedback effect of
testosterone on both the hypothalamus and the pituitary to regulate
gonadotropin secretion is well known. In humans, the negative feedback
of testosterone on the hypothalamo-pituitary axis to inhibit secretion
of both LH and FSH is the basis for the current approach to male
contraception (169). It is now becoming clear that a major component of
the negative feedback action of androgens on gonadotropin secretion is
mediated via aromatization to estrogen (170, 171, 172, 173). In particular,
studies in humans showed that administration of estradiol could further
enhance gonadotropin suppression that was induced by a
testosterone-based contraceptive (174), further indicating estrogen’s
role as a negative feedback regulator of gonadotropin secretion. The
demonstration of ER and ERß in the rodent hypothalamus and
pituitary (73, 142, 175, 176), of ERß in the monkey pituitary (143),
together with the immunolocalization of aromatase in the brain and in
particular the hypothalamus (177), indicates that estrogen has an
important role in these tissues.
During pubertal development in the rodent, numerous studies have shown that neonatal exposure to either estrogens or estrogen-like compounds promotes changes in gonadotropin secretion (178, 179, 180) and can, in fact, alter the organization of the hypothalamo-pituitary-testis axis so that changes persist into adulthood (181). A single high dose of estradiol benzoate to 1-day-old male rats causes a reduction in both GnRH secretion and pituitary responsiveness to GnRH (182), as well as the profound suppression of circulating FSH, LH and, consequently, testosterone levels (180). Interestingly, recent studies administering low doses of estrogenic compounds during the neonatal period in rats could actually stimulate serum FSH levels during puberty, an effect that could not be explained by changes in inhibin B (178). Given that the appropriate concentrations of LH and FSH, as well as a tightly regulated onset of secretion during the neonatal and pubertal periods, is fundamental to whether normal spermatogenesis proceeds (183, 184), neonatal estrogen exposure can have important long-term effects on the hypothalamo-pituitary-testis axis and thus spermatogenesis. Interestingly, whether or not estrogen administration to juvenile mice will interfere with the hypothalamo-pituitary-gonadal axis appears to be strain-dependent (185), which could lead to confusion when interpreting the literature on the interaction between estrogen, the regulation of pituitary hormone production, and fertility.
In the adult, there are many examples of a role for estrogen in the regulation of gonadotropin secretion. Adult male rats given increasing doses of estradiol for 10 days showed significant decreases in circulating concentrations of FSH and LH, which leads to subsequent reductions in serum and testicular testosterone levels (186). Surprisingly, the authors of this study also noted a stimulatory effect of low doses of estradiol on FSH, as was demonstrated by studies in the neonate (178), and in adult hypogonadal (hpg) mice given reasonably physiological doses of estradiol (187), indicating that estrogen can participate in both negative and positive effects on the pituitary in the male. Some of the effects of estrogen on FSH secretion may be mediated by its ability to promote changes in the in vitro Sertoli cell production of inhibin B (188), which is an important mediator of FSH secretion in males (Ref. 189 and references therein). Interestingly, estrogen has also been shown to increase the expression of the ßB-subunit of inhibin B in breast cancer cells (190). Therefore, estrogen may play a role in the regulation of this peptide that is primarily involved in mediating pituitary gonadotropin secretion, but may also have other roles in the testis (see Ref. 191 for review).
While estrogen administration clearly causes decreases in circulating gonadotropin levels, the administration of aromatase inhibitors causes increases in serum LH and testosterone in adult dogs (192) and rats (193), and in serum testosterone and the responsiveness of Leydig cells to a bolus injection of LH in monkeys (194). An increase in the circulating concentration of FSH was also seen in rats treated with an aromatase inhibitor (193). Administration of an aromatase inhibitor to men causes increases in the circulating concentrations of LH, FSH, and testosterone (172, 195). An interesting study in which the effects of administration of an aromatase inhibitor was compared in normal vs. GnRH-treated hypogonadotropic-hypogonadal men demonstrated that estrogen acts at the hypothalamus to decrease GnRH pulse frequency and at the pituitary to decrease responsiveness to GnRH (173). Thus, in the human, it is clear that estrogen is an important regulator of the hypothalamo- pituitary-gonadal axis in both sexes.
Investigation into the role of estrogen in the control of the
hypothalamo-pituitary-testis axis has been facilitated by transgenic
mouse models (see Table 2). As would be
predicted from aromatase inhibitor studies, male mice that lack a
functional aromatase gene (ArKO) have increased levels of serum LH and
testosterone yet normal levels of FSH (8, 196). Although ER is
present in the mouse hypothalamus and is the only ER in the mouse
pituitary (73), male mice deficient in ER (ERKO) showed
surprisingly little change in LH and FSH levels, although serum
testosterone levels were higher (5). A later study with a larger number
of animals showed that ERKO males had a 2-fold increase in both
circulating LH and in the pituitary content of LH ß mRNA, but
confirmed that ERKO animals had no change in FSH (197). An elegant
series of experiments performed by Lindzey and colleagues (197), and
reviewed in detail elsewhere (27), suggested that estrogen is likely to
be involved in facilitating the negative feedback effects of
testosterone on the male mouse hypothalamo-pituitary-testis axis and
further demonstrated that ERKO animals may have an increased
sensitivity to androgens, suggesting an altered organization of the
hypothalamo-pituitary-testis axis. The endocrine profile of the ERßKO
has not been detailed; however, the fact that ERßKO males are fully
fertile suggests no endocrine disruption (198).
|
B. Efferent ductules and epididymis
After the release of mature spermatids from the Sertoli cell
during spermiation, spermatids via the seminiferous tubule fluid
proceed into the rete testis. Arising from the rete testis are a series
of tubules known as the efferent ductules, which connect the rete
testis to the initial segment at the head of the epididymis (see Ref.
199 for review). The primary function of the efferent ductules is to
resorb water, ions, and proteins, and various studies have shown that
approximately 90% of rete testis fluid is resorbed within these
ductules, so that spermatozoa become concentrated as they enter the
epididymis, thereby ensuring that a large number of spermatozoa are
released upon ejaculation (Refs. 162, 199, 200, 201 and references
therein). The passage of sperm through the initial segment, caput,
corpus, and cauda epididymis is essential for the final maturation of
sperm. The epididymis synthesizes and secretes numerous proteins and
also actively endocytoses substances from the epididymal lumen to
remove secreted proteins that are no longer required (see Refs. 202, 203 for review). Thus, the resorption of fluid through the efferent
ductules, as well as the attainment of various morphological,
biochemical, and motile properties during passage through the
epididymis, is fundamental for adequate sperm content of the ejaculate
and for full fertilizing capabilities.
Whereas a role for androgen in the regulation of epididymal function is
well known (204, 205, 206, 207), studies in the 1970s showed that there is a high
concentration of estrogen-binding sites in the immature and adult
epididymis (208, 209, 210), suggesting a role for estrogens in sperm
maturation and male reproduction. The concentration of estrogen leaving
the testis is far higher than in the circulation (see Section
IV.D and Ref. 162 for review), and estrogen receptors (ER and
ß) are present in the efferent ductules and epididymis (73, 125, 134, 140, 159, 160, 211, 212, 213) (see Section IV). In the rat
efferent ductules, ER is expressed at 3.5 times the level seen in
uterus (159). The presence of abundant estrogen and ER in the efferent
ductules and epididymis points to a role for estrogen in the regulation
of these tissues and thus in modifying sperm maturation and function.
A study by McLachlan and colleagues in 1975 (24) showed that prenatal exposure of mice to diethylstilbestrol (DES) caused epididymal granulomas, suggesting an action of exogenous estrogen on epididymal development. Numerous studies since have shown that neonatal estrogen exposure causes an impairment of efferent ductule and epididymal development and function that can lead to deleterious effects on fertility in adulthood (e.g., Refs. 134, 214, 215, 216, 217, 218, 219, 220). While at least part of these effects could be attributed to changes in the hypothalamo-pituitary-gonadal axis, a direct action on the epididymis to mediate these changes is likely given the high levels of ER expression.
Although the above studies indicate that estrogen administration can
affect the development of efferent ductule and epididymal function and
lead to an impairment of male reproductive function, the first major
insights into the mechanism of estrogen action in the efferent ductules
and epididymis, and the requirement for estrogen in male fertility,
were gained from the ERKO mouse. The ERKO male mouse is
infertile, due to disruptions to spermatogenesis, reduced
epididymal sperm content, reduced sperm motility and fertilizing
ability, as well as defects in reproductive behavior (4, 5, 221) (see
Ref. 27 for review and Table 2 and Section V.G). Studies by
Hess and colleagues (6) clearly demonstrated that dysfunction of the
efferent ductules of the ERKO mice contributed to the impairment of
fertility. ERKO testis weights initially increase during
postpubertal testicular development, due to an accumulation of fluid in
the lumen of the seminiferous tubules (5, 6), suggesting either an
excess of fluid secreted by the testis and/or a failure of fluid
resorption by the efferent ductules. In vivo efferent
ductule ligation experiments showed that the first possibility did not
occur, since ERKO testes secreted less fluid than wild-type animals
(6). Isolated efferent ductules were then cultured and ligated at
either end in vitro to analyze the removal of luminal fluids
by the efferent ductule epithelium. While ligated wild-type ductules
efficiently remove luminal fluid and collapse in vitro, the
efferent ductules of the ERKO mice were shown to swell, indicating
that fluid resorption by the efferent ductules in the ERKO mice was
impaired (6). Interestingly, efferent ductules isolated from wild-type
animals treated in vivo with the dual ER antagonist ICI
182,780, did not swell to the same extent as ERKO tissues,
suggesting that estrogen acting via ERß may contribute to efferent
ductule function (6). There have been no reports of efferent ductule
and epididymal dysfunction in ERßKO mice that exhibit normal
fertility (27, 198). However, more recent studies in which the effects
of the dual ER/ERß antagonist ICI 182,780 on adult efferent
ductule function in vivo were compared with the phenotype of
ERKO animals showed that both ICI-treated and ERKO animals
displayed similar disruptions to the rete testis and efferent ductules,
suggesting that ER, rat
