From the Laboratory of Biochemistry and Metabolism, NIDDK, National Institutes of Health, Bethesda, Maryland 20892-1812
It has now been over 60 years since Riddle
et al. (1) purified a hormone from the anterior pituitary
gland, which stimulated milk secretion in the mammary gland of virgin
rabbits. They named it prolactin (PRL).1
Since then, the synergistic approaches of biochemistry, physiology, molecular biology, and cell biology have unveiled several molecular switches in the PRL signaling cascade (Fig. 1).
Loss-of-function studies in the mouse have now provided clear insight
into the biology of two components of the PRL pathway. A mandatory role for the prolactin receptor (PRLR) and for the signal transducer and
activator of transcription (Stat) 5a in mammopoiesis and lactogenesis was established (2, 3). Although Stat5a is in the line of fire of many
signals such as PRL, growth hormone (GH), and several cytokines, its
absence in vivo reveals an unexpected level of specificity.
PRL is a 23-kDa peptide, which is mainly synthesized in
lactotrophic cells of the anterior pituitary of vertebrates. Many functions have been attributed to this hormone. PRL regulates gonadal
functions (4) and behavior such as nest building and the retrieval of
offspring (5), and it exerts multiple effects on the immune system (6).
The best characterized role of prolactin, however, is its ability to
induce lobuloalveolar growth in the mammary gland (7) and to stimulate
postpartum lactogenesis. These properties are mediated through the
activation of genes involved in growth control and differentiation.
The cloning of the PRLR in 1988 (8) unleashed efforts to elucidate the
cascade of molecular switches linking the receptor with the target
genes. The PRLR is a single chain transmembrane protein that belongs to
the cytokine receptor superfamily and is expressed in a wide variety of
tissues. Alternative splice products of the primary transcript yield
PRLRs with short and long cytoplasmic tails. Both forms of the PRLR can
dimerize upon ligand binding and activate the Janus kinase 2 (Jak2),
Fyn, and mitogen-activated protein kinase to promote cell growth (9, 10). However, only the long form does activate transcription of the
In a landmark experiment in 1965, Topper and colleagues (12)
demonstrated that synergistic signaling by insulin, hydrocortisone, and
PRL is required to produce casein in organ explants from mammary tissue. Targeted efforts in the late 1970s and early 1980s led to the
cloning of milk protein genes (13-16). Thereafter cell lines derived
from mammary tissue and studies using transgenic animals were
instrumental in the identification, characterization, and biological
verification of PRL response elements in the promoters of genes
expressed specifically in the mammary gland (17-22). By 1988 it was
clear that promoter sequences from the genes encoding the whey acidic
protein (WAP) and PRL, GH, epidermal growth factor, erythropoietin, and many
cytokines use STAT proteins to regulate the transcription of specific genes through the JAK-STAT pathway (10). Ligand binding triggers dimerization or oligomerization of receptors. Receptor-associated tyrosine kinases (JAKs) cross-phosphorylate each other as well as
the tyrosine residues on the receptors. Subsequently,
SH2-containing latent cytoplasmic proteins from the STAT family are
recruited to the receptor complex and phosphorylated by the JAKs. Two
STAT proteins dimerize, translocate into the nucleus, and activate gene
transcription by binding to GAS in gene promoters. The ability of
individual receptors (cytokines, GH, PRL) to activate overlapping but
distinct sets of homo- and heterodimerizing STAT proteins is thought to
contribute to their signal specificity. For example, interferon Functional differentiation of the mammary gland is a crucial step
in the reproductive cycle of mammals (28). The development of the gland
proceeds in distinct phases. In newborn mice a rudimentary system of
small ducts is present, which grows slowly until the onset of puberty
when pronounced ductal growth occurs. Development of the ducts
continues in cycling virgins leading to the formation of a ductal tree,
which fills the entire mammary fat pad. Terminal differentiation of
alveolar epithelial cells is completed at the end of gestation with the
onset of milk secretion at parturition. Distinct steps of cellular
differentiation take place during this process, which are defined by
the sequential activation of genes coding for milk proteins (29).
Stricker and Grueter (30) demonstrated in 1928 that pituitary extract
injected into castrated virgin rabbits induced milk secretion; this was
a harbinger that PRL is essential for mammary function. However, the
presence of lactogenic hormones is not sufficient to explain the
complex development and temporal regulation of gene expression observed
in the emerging gland during pregnancy. Estrogen and progesterone are
required for ductal outgrowth (31) and alveolar proliferation (32), respectively. In addition, growth regulators and cell cycle progression are obligatory for alveolar proliferation and differentiation.
PRL is a central player in the cast of characters. Alveolar
proliferation and the induction of terminal differentiation, as indicated by the transcriptional activation of milk protein genes, require the presence of prolactin (20, 28, 33, 34). In the functional
postpartum gland high levels of activated Stat5 can be found while only
small amounts of phosphorylated Stat1 and Stat3 have been detected
(35). Phosphorylation of Stat5a and -5b is very low in mammary tissue
of virgins and during early pregnancy but rises sharply after day 14 of
pregnancy. This led to the hypothesis that the activation of Stat5 is a
critical step in the terminal differentiation of mammary secretory
epithelium (35).
Genetic Disruptions of the PRL Pathway Evidence that PRL signaling induces mammopoiesis and lactogenesis
comes largely from studies with organ cultures and cell lines.
Similarly, our knowledge of the roles of Stat5 in PRL, GH, and cytokine
signaling comes primarily from tissue culture cells. Nonetheless, these
observations fall short of uncovering the roles of PRL and Stat5a and
-5b in normal physiological processes in vivo. The
generation of mice, from which the genes encoding the PRLR, Stat5a, and
Stat5b2,3 have been inactivated, has now provided clear insight
into their biology in vivo (2, 3). Since the PRLR is
expressed in many organs of the developing fetus and Stat5s are in the
line of fire of multiple signaling pathways, late fetal or neonatal
lethality could have been expected in the absence of these proteins. In contrast, mice deficient in PRLR or Stat5a or Stat5b were born and
survived until adulthood. Defects were confined to a few tissues and
specific physiological conditions.
Females with only one intact PRLR allele failed to lactate after their
first pregnancy due to greatly reduced mammary development (3). Clearly
epithelial cell proliferation during pregnancy depends on a threshold
of PRLR, which cannot be obtained with just one functional allele.
However, mammary gland development after the second pregnancy was
sufficient for a successful lactation, demonstrating that continued
hormonal stimuli will eventually lead to functional development. It is
currently not clear whether other growth pathways will compensate for a
decreased PRLR population. In contrast, the progesterone (32) and
estrogen (31) receptors, although essential for mammopoiesis, mediate
functional development in the presence of only one functional
allele.
The only noticeable phenotype of Stat5a-deficient mice is their
inability to lactate due to a failure of the gland to develop and
differentiate during pregnancy (2). Whether this is due to a unique
property of Stat5a relative to Stat5b is unclear. Stat5a and Stat5b
exhibit superimposable expression patterns during mammary gland
development (35) and have a 96% similarity, although the majority of
the differences lie in the transcriptional activation domain at the
carboxyl termini (27, 36). The observation that mice heterozygous for
the PRLR deficiency exhibit a comparable phenotype suggests that a 50%
reduction in signaling capability can cause a lactation deficiency. In
addition, the disruption of the Stat5a locus also clearly had an effect
on the amount of phosphorylated Stat5b (2). Whether the Stat5 proteins
have unique functions will also be addressed by the derivation of
Stat5b-deficient mice. Two groups2,3
have recently obtained
such mice although the preliminary results appear different. In one
case2 the phenotype of the Stat5b-deficient mice is distinct
from that of the Stat5a-deficient mice. These mice exhibit reduced
growth of males and a severely compromised fertility in females.
Interestingly, the reproductive lesions in the Stat5b-deficient mice
were similar to those observed in the PRLR-deficient mice. The
Stat5b-deficient males were characterized by a decrease in body growth
profile to the slower rate of wild type females.2 This growth
defect first emerges at puberty and is apparently due to a loss of
responsiveness to plasma growth hormone pulses, proposed to be a
male-specific Stat5b-mediated signaling pathway in rodents. A second
group3 obtained a phenotype that was comparable with the
Stat5a-deficient phenotype in exhibiting only a lactation deficiency.
The basis for the difference in phenotypes is currently unclear
although it is essential to characterize multiple, independently
derived strains of mice. It will also be important to obtain mice that are deficient in both genes. Such mice have recently been
generated,3 and the phenotype of these mice should be reported
soon. Although both the WAP and It is intriguing to speculate why widespread and general signaling
cascades involving the PRLR and Stat5a are critical for mammopoiesis
and lactogenesis, while their requirements for the development of other
tissues appear to be far less stringent. Since the mammary gland is a
recent acquisition in the phylogenetic scale of organ evolution, it may
well be possible that redundant signaling pathways have not been
developed.
Future progress in understanding the cell- and ligand-specific
effects of signaling molecules with a widespread distribution, such as
the PRLR and the transcription factors Stat5a and Stat5b, will depend
on our ability to further manipulate the genetic components of the
mouse. Aberrant mammary development and function observed in the
absence of Stat5a and the PRLR are likely the result of accumulated
consequences of misguided signaling during puberty and pregnancy. To
understand the molecular basis of these compound lesions, it will be
necessary to identify the cell type and record the time point at which
the initial impact of aberrant signaling occurs.
Experimental approaches to identify cell types and developmentally
crucial time windows are becoming available. These include the
time-sensitive activation of genes (38, 39), and the cell- and
development-specific inactivation of genes by site-specific recombination, such as the Cre-lox technique (40). The time-sensitive activation of a Stat5a transgene in a Stat5a-deficient background could
help in the identification of developmental steps during puberty and
pregnancy. This could be accomplished either with the classical
tetracycline-responsive system (39) or by
adenovirus-mediated gene transfer into mammary epithelial cells.
Alternatively, the combination of the Cre-lox and the
tetracycline-responsive expression system will permit the deletion of
the Stat5 gene in a tissue- and temporal-specific fashion (40).
The deletion of more than one gene in the PRL pathway gene will provide
further insight into both the specificity and redundancy of PRL
switches. Considering the variety of signals that utilize Stat5a and
-5b, it is likely that the simultaneous deletion of both genes will
have complex consequences. The tools of tissue- and temporal-specific
deletions of genes will find an application. Finally, the mammary gland
lends itself to transplantation studies, which permits the analysis of
wild-type epithelium in a mutant host and vice versa. Such experiments
allow the identification of the tissue affected by elimination of one
component in the pathway.
The past 60 years of PRL research were characterized by a synergy of
biochemistry, genetics, and physiology, which culminated in the
identification of molecular switches in the PRL signaling pathway.
Future research will focus on how these switches control the physiology
of the mammary gland.
Fig. 1.
Prolactin signaling in the mammary
gland. PRL binds to its receptor and causes the PRLR to dimerize.
Receptor-associated tyrosine kinase Jak2 phosphorylates the prolactin
receptor and the signal transducers and activators of transcription
Stat5a and Stat5b. Activated Stat5a and -5b are transported into the nucleus, bind to GAS sequences (TTCNNNGAA), and induce transcription of
target genes that promote proliferation, differentiation, and lactogenesis.
[View Larger Version of this Image (34K GIF file)]
-casein gene in transfected tissue culture cells (11). Jak2 in turn
phosphorylates specific tyrosine residues in Stat1, -3 and -5 (9, 10).
These activated STAT proteins bind to and induce transcription from
promoters containing
-interferon activation sites (GAS) (TTCNNNGAA)
(10).
-lactoglobulin contain sufficient genetic
information to target transcription exclusively to mammary tissue and
to respond to PRL signals (23, 24). In the quest for the identification
of PRL response elements, a sequence (GAS) in the promoter of the
-casein gene was identified, which was specifically recognized by a
phosphorylated nuclear protein from mammary tissue. This protein was
named mammary gland factor (MGF) (25). Transfection experiments of
mutated promoter fragments into HC11 cells led to the demonstration
that GAS (TTCNNNGAA) convey PRL responsiveness and are recognized by
MGF (25). Experiments using transgenic mice demonstrated that GAS in
the promoters of the genes encoding WAP (17) and
-lactoglobulin (21,
22) are critical for maximal gene activity and PRL response in
vivo. MGF by itself, however, is not sufficient for optimal
activity, and it may cooperate with juxtaposed transcription factors,
such as nuclear factor 1 (17). MGF was cloned in 1994 from sheep mammary tissue and recognized as a new member of the family of STAT
proteins (26). In the mouse Stat5 exists as two isoforms (5a and 5b)
with a 96% similarity (27).
activates Stats 1, 2, and 3 and exerts a growth-retarding effect. PRL,
in contrast, activates Stats 1, 3, and 5 in many cell lines. It induces
transcription from the promoter of the
-casein gene in mammary
epithelial cells and causes proliferation of Nb2 lymphoid cells. Stat5a
and Stat5b are expressed in most, if not all, tissues, and they can be
activated in tissue culture cells by PRL, GH, epidermal growth factor,
and many cytokines (9). This suggests that Stat5 transcription factors
are components of different signaling pathways leading to cell growth
and differentiation. In the mammary gland, Stat5a and -5b are activated
by PRL and probably placental lactogen.
-casein genes contain GAS, only WAP
gene expression was reduced in Stat5a-deficient mice, suggesting that
-casein gene expression is primarily controlled by other
transcription factors. Finally, mice with an inactive PRL gene have
been generated,4 and the physiological
consequences should be known soon.