Molecular Characterization of Prostate Hyperplasia in Prolactin Transgenic Mice

(1)

Molecular Characterization of Prostate Hyperplasia in Prolactin Transgenic Mice

Karin Dillner

Department of Physiology and Pharmacology Sahlgrenska Academy, Göteborgs University

Sweden 2003

(2)

All previously published papers were reproduced with permission from the publishers.

Printed by Svenska Tryckpoolen AB

(3)

- 1 -

ABSTRACT

Benign prostatic hyperplasia (BPH) and prostate cancer are age-related diseases, affecting a majority of elderly men in the western world, and are known to be influenced by several different hormones, including sex hormones. Although the hormone prolactin (PRL) is well known to exert trophic effects on prostate cells, its involvement in the pathophysiology is still poorly characterized. In order to evaluate the potential role of PRL in promoting prostate growth, we used PRL-transgenic mouse models that develop prostate phenotypes.

The Mt-PRL transgenic mouse model, ubiquitously overexpressing the rat PRL transgene, develops a dramatic prostate hyperplasia with concurrent chronic hyperprolactinemia and elevated serum androgen levels. In a castration and androgen- resubstitution study, we demonstrated that supraphysiological serum androgen levels are not required for the progress of prostate hyperplasia in adult Mt-PRL transgenic mice.

Furthermore, androgen treatment does not induce prostate hyperplasia in wildtype mice. To address the role of local PRL action in the prostate, a new transgenic mouse model (Pb- PRL) was generated using the prostate-specific probasin minimal promoter to drive expression of the rat PRL gene. The androgen-dependency of the probasin promoter resulted in onset of the PRL transgene expression at puberty. The Pb-PRL transgenic mice also develop a significant prostate hyperplasia, evident from 10 weeks of age and the hyperplasia increases with age. In contrast to the Mt-PRL transgenic mice, the Pb-PRL transgenic mice display normophysiological serum androgens levels throughout animal life span. The prostates of both the Mt- and Pb-PRL transgenic mice display a prominent stromal hyperplasia with mild epithelial dysplastic features, leading to an increased stromal/epithelial ratio. Accumulation of secretory material is also a major characteristic.

Immunohistochemistry analysis of both the PRL transgenic models’ prostates showed an increased androgen receptor distribution in both the epithelial and stromal cells.

Microdissections demonstrated an increased ductal morphogenesis in the Mt-PRL prostate compared to Pb-PRL and controls, indicating that PRL stimulates, directly or indirectly via increased androgen action, prostate ductal morphogenesis in the developing prostate gland.

The use of differential gene expression technologies enabled characterization of the molecular mechanisms involved in the prostate hyperplasia. Of particular interest is the potential significance of reduced apoptosis for the development/progression of the prostate phenotype. This finding was further confirmed by immunohistochemical analysis using two different apoptosis markers. Moreover, in line with the prominent expansion of the stromal compartment, were the identified changes in gene expression seen in the PRL transgenic prostate, suggesting that activation of the stroma is important for the development of the prostate hyperplasia.

Altogether, there are histological and molecular similarities between the prostate hyperplasia of PRL-transgenic mice and human prostate pathology, including both BPH and prostate cancer.

Key words: Prolactin-transgenic, mouse, prolactin, prostate hyperplasia, gene expression analysis

(4)

- 2 -

ORIGINAL PAPERS

This thesis is based on the following papers, which are referred to in the text by their Roman numbers (I-IV);

I. Kindblom J, Dillner K, Ling C, Törnell J and Wennbo H

Progressive Prostate Hyperplasia in Adult Prolactin Transgenic Mice is Not Dependent on Elevated Androgen Serum Levels.

Prostate 2002 Sep 15;53(1):24-33.

II. Dillner K, Kindblom J, Flores-Morales A, Pang ST, Törnell J, Wennbo H and Norstedt G

Molecular Characterization of Prostate Hyperplasia in Prolactin- Transgenic Mice Using cDNA Representational Difference Analysis.

Prostate 2002 Jul 1;52(2):139-49.

III. Kindblom J, Dillner K, Sahlin L, Robertson F, Ormandy CJ, Törnell J and Wennbo H

Prostate Hyperplasia in a Transgenic Mouse with Prostate-Specific Expression of Prolactin

Endocrinology, 2003, in press.

IV. Dillner K, Kindblom J, Flores-Morales A, Shao R, Törnell T, Norstedt G and Wennbo H

Gene Expression Analysis of Prostate Hyperplasia In Mice Overexpressing the Prolactin Gene Specifically in the Prostate.

Submitted for publication.

(7)

- 5 -

LIST OF ABBREVIATIONS

-/- aa AP AR BPH cDNA Cy DLP DP DP1, 2, 3 ECM ER EST FDR GH hPRL LP MMP mRNA Mt-1 Mt-PRL Pb

Pb-PRL PIF PIN

PL PRL PRLR PSA RDA rPRL RT-PCR SAM ssDNA Stat SMA TUNEL TURP UGM UGS UTR VP

homozygous gene-deficiency amino acids

anterior prostate androgen receptors

benign prostatic hyperplasia

complementary deoxyribonucleic acid cyanine

dorsolateral prostate dorsal prostate

difference products 1, 2, 3 extracellular matrix estrogen receptor

expressed sequence tags false discovery rate growth hormone human PRL lateral prostate

matrix metalloproteinase messenger ribonucleic acid metallothionein-1 gene

The metallothionein-1 promoter - rat prolactin gene probasin gene

The minimal probasin promoter - rat prolactin gene prolactin inhibiting factors

Prostatic intra-epithelial neoplasia placental lactogen

prolactin

prolactin receptor

prostate specific antigen

representational difference analysis rat prolactin

reverse transcription polymerase chain reaction Significance Analysis of Microarrays

single stranded DNA

signal transducers and activators of transcription Statistics of Microarrays Analysis

terminal deoxynucleotidyl transferase dUTP nick end labeling transurethral resection of the prostate

urogenital mesenchyme urogenital sinus

untranslated region ventral prostate

(8)

- 6 -

INTRODUCTION

THE PROSTATE GLAND

The prostate gland is an exocrine gland that is found only in mammals. The main function of the gland is to produce a major fraction of the seminal fluid, including enzymes, amines, lipids and metal ions. One unique function of the prostate gland is the capacity to produce, accumulate and secrete high levels of citrate [1]. The prostate varies in its anatomy, biochemistry and pathology between different species. The mature mammalian prostate is a glandular organ consisting of epithelial and stromal cell types that are hormonally regulated. The epithelium consists of a single layer of polarized columnar epithelial cells together with basal and neuroendocrine cells. The epithelial cells supply secretions that empty through ducts into the urethra to form the major component of the seminal plasma of the ejaculate. The surrounding stromal compartment comprises of fibroblasts, smooth muscle cells and loose collagenous extracellular matrix (ECM), in addition to neuronal, lymphatic and vascular components.

Interest in understanding the biology of the prostate has largely been driven by the high incidence of prostate diseases, including benign prostatic hyperplasia (BPH) and prostate cancer.

P

ROSTATE DEVELOPMENT

The development of the male reproductive tract is dependent upon androgens and mesenchymal-epithelial interactions [2]. The initial event in prostatic morphogenesis is the outgrowth of solid cords of epithelial cells, so-called prostatic buds, from the urogenital sinus epithelium into the surrounding urogenital sinus mesenchyme. In rodents, this occurs in a precise spatial pattern that establishes the lobar subdivisions of the prostate [2, 3]. In rodents, the critical time period for ductal budding and the consequent process of ductal growth and branching initiate around day 15 of gestation and conclude approximately 4-5 weeks postpartum [4-6]. The branching morphogenesis is almost entirely complete by 2 weeks after birth in the mouse [4]. At this time, serum testosterone levels are still low and the increase in prostatic wet weight is modest. As shown by neonatal castration studies, the neonatal prostatic ductal morphogenesis is sensitive to, but does not require, chronic androgen stimulation [7]. The prepubetal growth of the prostatic ductal network is considered non-uniform, where the growth is

(9)

- 7 -

highest in the distal region, at the ductal tips, and much lower in the proximal region closest to the urethra [8, 9]. At puberty, the testosterone levels raise significantly, and the rodent prostatic wet weight and DNA content increase more rapidly [7]. In contrast, the human prostate morphogenesis occurs entirely during the fetal period, with ductal development primarily occurring in the first half of gestation [10].

P

ROSTATE ANATOMY AND STRUCTURE IN HUMAN AND RODENTS

There are fundamental differences between the prostate anatomy in human, dog and other primates and non-primates, e.g. rodents. The human prostate is associated with the urethra contiguously below the urinary bladder and prostatic ducts emanate from the urethra radiating towards the periphery completely surrounding the urethra. The adult human gland can be divided into four zones based on morphology; the anterior fibromuscular stroma, the central zone, the peripheral zone and the transition zone. The two latter are of more clinical interest because prostatic carcinoma arises nearly exclusively from the peripheral zone and BPH from the transition zone [11].

In contrast, the process of branching morphogenesis in rodents ultimately gives rise to three distinct bilaterally symmetrical prostatic lobes: the anterior prostate (AP; also known as the coagulating gland), the dorsolateral prostate (DLP), and the ventral prostate (VP). The DLP is sometimes further divided into the dorsal prostate (DP) and the lateral prostate (LP). Individual lobes are located in specific positions around the urethra, but not completely circumscribing it [2]. This explains why rodents, in contrast to most humans, do not suffer from urinary tract symptoms following prostate enlargement.

The ducts of each of the rodent prostatic lobes have a characteristic branching pattern [4]. The VP and LP lobes are attached to the urethra by two or three main ducts that show extensive so-called “oak tree” branching, whereas the DP lobe demonstrates multiple main urethral ducts with less extensive so-called “palm tree” branching morphology [4]. Furthermore, the ductal system also shows regional variation in morphology and functional activity [12] and therefore ductal system of each lobe can be further subdivided into regional segments, defined as proximal, intermediate and distal with respect to the urethra [13]. The VP has no clear homologous counterpart in the prostate of higher animals, whereas the DLP are considered the most homologous to the human prostate [14, 15].

The prostate tissue can be divided into epithelial and stromal parts and the proportion between epithelial and stromal compartments differs between species. In the adult rat the stromal:epithelial ratio is 1:5, whereas in humans,

(10)

- 8 -

the stromal and epithelial cells are present in approximately the same number in the normal prostate [16, 17].

Another important species difference between rodent and human prostate is the presence of the androgen-regulated serine protease, prostate specific antigen (PSA) in human. PSA is produced by both prostate epithelial cells and prostate cancer cells and is the most commonly used serum marker for prostate cancer as well as to monitor responses to therapy. Genes related to human PSA have been detected in several non-human primate species, but not in other mammalian species, including rodents [18].

PROSTATE DISORDERS

Prostate gland disorders are age-related diseases affecting a majority of elderly men in the western world. Among mammals with a prostate gland, humans and dogs are the only species known to suffer from BPH and prostate carcinoma [19].

B

ENIGN

P

ROSTATIC

H

YPERPLASIA

BPH is characterized as a slow, progressive enlargement of the prostate gland, which eventually causes obstruction and subsequent problems with urination.

However, BPH is believed to be neither a premalignant lesion nor a precursor of prostate cancer. The incidence of BPH is increasing dramatically with age from about 50% at 50 year of age to 90 % by the ninth decade of life [20]. The BPH progression is characterized by hyperplasia of both the stromal and epithelial compartments. When calculating the stromal:epithelial ratio, clinical reports have firmly established a dominance of the stromal compartment in BPH tissues, which is in contrast to the balanced epithelial and stromal distribution in normal prostate tissue [21-24]. Furthermore, in symptomatic BPH patients the stromal:epithelial ratio has been reported to be significantly higher than in asymptomatic patients [22].

Testosterone is the principal circulating androgen. In men, it is secreted primarily by the testis, with the adrenal glands providing a minor contribution. To be maximally active in the prostate, testosterone must first be converted to dihydroxy testosterone (DHT) by the enzyme 5-alpha reductase. DHT is about five times more potent as an androgen within cells than testosterone, and it binds readily to the androgen receptors (AR) in the nucleus. Androgens are clearly required for development of BPH and

(11)

- 9 -

reduction of androgenic effects through 5-alpha reductase inhibitors is utilized in the pharmacotherapy of BPH. Treatment with 5-alpha reductase inhibitors rapidly reduces DHT serum levels and over time results in an average decrease in prostate volume [25]. In addition, alpha-1 adrenoreceptor antagonists are increasingly used, either given in combination with a 5-alpha reductase inhibitor or separately [26]. The mechanism of action of the alpha- 1 adrenoreceptor antagonists is primarily to reduce the contractility of the smooth muscle cells in the bladder neck and prostatic urethra which result in an improved urinary flow. The traditional surgical techniques such as transurethral resection of the prostate are still appropriate for some patients, although with improved medical treatments now available, the number of men undergoing surgery is most likely declining [27].

Possible theories of BPH etiology

Despite of BPH’s obvious importance as a major health problem, little is known in terms of the biological processes that contribute to the pathogenesis of BPH. However, a number of theories have been proposed over recent years to explain the etiology of the pathological phase of BPH and the most typical will be described briefly below. Although they may show some degree of contradictions, they most likely contribute together to the pathogenesis of BPH.

One of the theories, the dihydroxy testosterone theory, was originally based on the failure of BPH to develop in men castrated prior to puberty. Although controversy still exists, a decreased testosterone/DHT ratio, due to both decrease in plasma testosterone levels and possibly an increase in DHT levels, in elderly men with BPH, may be involved in the etiology of BPH [28, 29]. DHT levels in BPH may be higher than in normal prostate tissue.

The local levels of DHT may be increased by age, testicular endocrine function declines steadily with age and at 75 years of age, mean plasma testosterone levels are reported to be around 65% of levels in young males [30] and the decrease in bio-active (non sex hormone-binding globulin (SHBG)-bound) testosterone levels is even more pronounced [31]. This is likely due in part to the recognized increase in SHBG binding capacity associated with ageing. [32-34]. The DHT theory proposes that there is a shift in prostatic androgen metabolism that occurs with aging, which leads to an abnormal accumulation of the more potent DHT in the prostate, thus producing the enlarged prostate. Although the level of DHT in BPH tissue might not be elevated compared to normal tissue, it is very likely that the 5- alpha-reductase activity and AR levels are greater in BPH tissue than in controls. It is the binding of DHT to the AR which is important in

(12)

- 10 -

stimulating cell proliferation, and prostatic cells may therefore gradually become more and more sensitive to androgens with ageing [35]. Moreover, the reduction in prostate size upon suppression of androgen-mediated action, either by blocking secretion of circulating testosterone and adrenal androgen, inhibiting 5 alpha-reductase to prevent DHT formation, or blocking DHT binding to AR, have further proved the DHT theory [25, 36].

Another theory, the embryonic reawakening theory, was originally based on BPH histopathological features from which McNeal concluded that the prostate stroma undergoes an “embryonic reawakening”, resulting in inductive effects of the local stroma, which in turn induces hyperplastic changes in the epithelium through stromal-epithelial interactions. Somehow a shift of stromal-epithelial interactions occurs with aging, which leads to the inductive effect on prostatic growth.

A further theory is the stem cell theory [37]. The stem cell is a proliferative cell but the number of these cells within the prostate is unknown but believed to be very low. The normal behavior of stem cells include: (i) relatively undifferentiated; (ii) their numbers are preserved; (iii) unlimited proliferating potential; (iv) easily adapt to the environment; and (v) finally, but maybe the most important, they are pluripotent, which means that they can give rise to a number of different cell types. According to this theory, BPH could occur as a result from changed properties of the stem cells giving rise to a clonal expansion of cell populations [38].

One more theory, the estrogen-androgen imbalance theory, suggests that an age-associated imbalance between circulating estrogens and testosterone plays a role in the pathogenesis of BPH [39]. In humans, the serum testosterone and free-testosterone levels decrease with age, but the serum estradiol level is constant throughout life. Therefore, with age, creating an estrogen-dominant status compared to that at younger ages. These endocrine changes at mid-life have been extensively investigated through the past 30 years, and are commonly referred to as the “andropause” [40]. This results in a gradual, but significant, increase in the ratio of estradiol/testosterone in the serum [41].

Estrogen plays an important role in prostate pathophysiology (for more information, see section “ACTION OF ESTROGENS IN THE PROSTATE”).

An additional theory, the reduced apoptosis theory, suggests that a reduced rate of apoptosis is involved in the etiology of BPH [42], based on the observations of reduced apoptotic activity in BPH tissue compared to control [43, 44]. A homeostasis appears to exist after the prostate has reached its adult size, whereby the rates of prostatic cell growth and prostatic apoptosis are in equilibrium. This ensures that neither involution nor overgrowth takes

(13)

- 11 -

place, so that prostate size is constant. The reduced apoptosis theory suggests that the increased prostate volume in BPH is a function of a decrease in the rate of cell death perhaps in parallel with an increase in cell proliferation.

P

REMALIGNANT LESIONS OF THE PROSTATE

Prostatic intra-epithelial neoplasia (PIN) is associated with various alterations in prostatic cellular architecture such as dysplastic foci present in the prostatic ducts and acini [45]. Other histological or biological changes that have been reported include: decreased secretory differentiation, nuclear and nucleolar abnormalities, neovascularity, increased proliferative potential and genetic instability with variation of DNA content. Based on the morphological features, PIN can be divided into low and high grades. PIN is most commonly found in the peripheral zone of the human prostate. Genetic events in PIN have been linked to the development of prostatic carcinoma.

However, detailed analysis of the genetic alterations in PIN and matched cancer samples has been limited by the small size of foci of PIN, as well as by the marked morphologic heterogeneity and multi-focality of both lesions [46, 47]. Although, it seems like high-grade PIN is a precursor lesion to prostate carcinoma, the lack of adjacent high-grade PIN in many early cancers indicates the contradictory.

P

ROSTATE CARCINOMA

Prostate carcinoma remains one of the most common malignant diseases and is a leading cause of cancer-related deaths among men in the industrialized world. However, the vast majority of men harboring pathologic evidence of prostate cancer are not clinically diagnosed with this disease and it is far more common to die with prostate cancer than as a direct result of the disease. The development of new capillary blood vessels (angiogenesis) might well be one of the first steps in cancer progression. This may be induced by the abnormal tumor expression of growth factors, such as vascular endothelial growth factor (VEGF) and basic fibroblast growth factor (FGF) [48]. Further tumor progression and eventual metastasis may result from the fact that malignant cells are less adhesive to one another than normal cells. Cadherins are cell surface glycoproteins that are required for cell adhesion. Changes in the gene which controls cadherins could well be involved in progression and metastasis [49]. Extension of the tumor into the ECM is probably a complex alteration involving mediators between the malignant cells and the adhesive proteins of the ECM, e.g. integrins and fibronectin [50].

(14)

- 12 - Possible theories of prostate cancer etiology

The exact cause of prostate cancer is unknown, and part of the problem is the variability and heterogeneity of the tumor within the prostate gland.

However, the most established risk factors for development of prostate cancer include ageing, race, diet and a family history of prostate cancer. In addition, a number of theories for its pathogenesis have been suggested over recent years and together these theories most likely contribute to development of the disease.

One theory suggests an imbalance in growth regulation in the prostate. As in BPH, stromal-epithelial interactions and growth factors may also play a role in the pathogenesis of malignant disease of the prostate. These important local regulatory factors are involved in a balance which controls, not only cell growth, but also apoptosis. Inappropriate regulation of growth factors, which are produced not only by the target cells themselves, but also by neighboring cells, could develop a significant imbalance which, if prolonged, would be an important step in the genesis of the cascade of events which ultimately leads to prostate cancer.

Another possible theory proposes that the stroma undergoes an activation process, resulting in a formation of a so-called reactive stroma. There are considerable evidence that neoplastic stroma is different from the stroma of normal tissue. In an effort to maintain tissue homeostasis, the stromal compartment reacts to tumorigenic epithelium in a process similar to the generation of granulation tissue in wound repair stroma [51]. This activation of the stroma, resulting in a so-called “reactive stroma” and includes phenotypical changes of the stroma cells to a more myofibroblast-like phenotype (transient form between fibroblasts and smooth-muscle cells). The formation of reactive stroma is known to occur in many human cancers, including prostate, and is likely to promote tumorigenesis [52]. Furthermore, it is characterized by ECM remodeling, elevated protease activity, increased angiogenesis and an influx of inflammatory cells.

An additional theory involves the possibility of genetic instability in the growing tumor. This genetic instability refers to accumulation of several genetic defects that can occur either at the nucleotide level (e.g., insertion, deletion, or base substitution) or at the chromosomal level (such as, loss or gain of an entire chromosome or small portions) [53]. The genetic instability may result in the stimulation of proto-oncogene and/or inactivation of tumor suppressor genes. Carcinogenesis may develop when the genetic restraint and control in the growth of the cell is lost. Abnormal intracellular behavior can be

(15)

- 13 -

induced by oncogene activation or by a change in activity or character of the tumor suppressor genes. Proto-oncogenes are normal cellular genes involved in the regulation of growth and cellular differentiation, for example c-ras and c-myc. The simultaneous activation of these oncogenes could override the inhibitory restraints of neighboring cells and allow tumor proliferation. In parallel with oncogenes, normal cell also contain genes which protect against cancer, so-called tumor suppressor genes, for example the p53 and retinoblastoma (Rb) genes. The normal role of these genes include control of cell division, cell cycle check points and DNA repair, all to reduce and control the proliferation activity of the cells. It is known that loss of these genes may result in cancer, and it seems probable that the prostate tumors that occur in younger men, which appear to have a familial basis, may also be the result of specific gene deletions [53]. Furthermore, it is suggested that increased genetic instability is associated with decreased androgen-responsiveness and progressive behavior of human prostate tumors. Changes may take place which allow the development of androgen-insensitive cells and the death of androgen-sensitive cells. This would provide a further movement away from the modulating influence of androgens on the growth factors associated with normal cell regulation. However, it remains unclear whether this genomic instability is causing the progression of cancer or is the consequence of cancer [53].

PROLACTIN

Prolactin (PRL) has classically been considered as a pituitary-derived peptide hormone but over the last decade expression of the PRL gene has also been demonstrated in several extrapituitary tissues [54]. More than 70 years ago, PRL was found to be a pituitary factor that stimulates mammary gland development and lactation in rabbits, but since then PRL has been demonstrated to regulate more than 300 different biological functions, including reproduction, lactation growth, development, metabolism, immunomodulation, osmoregulation and behaviour [55].

G

ENE

,

STRUCTURE

,

AND VARIANTS

PRL is a member of the PRL/PL/GH hormone family, to which among others growth hormone (GH) and placental lactogen (PL) also belong to.

They all share genomical, structural, biological and immunological features [56, 57]. More recently, this family has been linked to a still more extended family of proteins, referred to as hematopoietic cytokines [58].

(16)

- 14 -

The gene encoding PRL is unique and is found in all vertebrates [55]. The rat PRL (rPRL) gene is located on chromosome 17, is approximately 10 kb long and composed of five exons and four introns. The human PRL (hPRL) gene is also approximately 10 kb, but located on chromosome 6 and contains an additional exon at the 5'-end [59]. This extra exon is only transcribed in extrapituitary sites, generating a 134 bp longer transcript differing in the 5´- untranslated region, compared to the pituitary transcripts [54]. The mature form of the protein contains 199 residues (23 kDa) and is folded into an all- α-helix protein. Although the tertiary structure has not been determined, PRL is predicted to adopt the four-helix bundle folding described for the GHs [55, 60].

Extrapituitary PRL protein is identical to pituitary PRL, but different promoters drive the expression of PRL in pituitary and extrapituitary sites in humans [61]. Pituitary PRL is controlled by a proximal promoter, which requires the Pit-1 transcription factor for trans-activation. In human, the promoter is divided into a proximal region and a distal enhancer, both of which are necessary for optimal pituitary-specific expression. The pituitary- type promoter and its regulation by dopamine, estrogens, neuropeptides and some growth factors have been well characterized [58]. In contrast, the synthesis of extrapituitary PRL is driven by a superdistal promoter, located 5.8 kb upstream of the pituitary start site. This promoter is silenced in the pituitary, does not bind Pit-1 and is not affected by dopamine or estrogens [60]. The superdistal promoter contains binding sites for several transcription factors but its regulation is poorly understood [62].

The PRL isoform 16K, was discovered more than 20 years ago as the N- terminal 16-kDa fragment resulting from the proteolysis of rat PRL by acidified mammary extracts [63]. The protease responsible for the cleavage of rat PRL into 16K PRL was identified as cathepsin D, whose implication in tumor progression is relevant [64]. 16K PRL was shown to have lost PRLR binding ability but otherwise to have acquired the ability to specifically bind another membrane receptor [65] through which it exerts anti-angiogenic activity [66]. Although this receptor is still not identified, some of its downstream signaling targets have been elucidated [67-70].

Moreover, a PRL-related hormone called proliferin (also known as mitogen- regulated protein (MRP)) [71] has been identified as a growth factor- inducible gene in immortalized mouse fibroblasts [72, 73], but in vivo it is produced primarily by the trophoblast giant cells [74]. Interestingly, reactivation of the proliferin gene expression has been associated with increased angiogenesis, as shown in a cell culture model of fibrosarcoma tumor progression [75].

(17)

- 15 -

C

ONTROL OF PROLACTIN SYNTHESIS

,

SECRETION AND REGULATION In contrast to what is seen with all the other pituitary hormones, the hypothalamus tonically suppresses PRL secretion from the pituitary. If the pituitary stalk is cut, PRL secretion increases, while secretion of all the other pituitary hormones falls dramatically due to loss of hypothalamic releasing hormones. Dopamine serves as the major inhibiting factor on PRL secretion.

Dopamine is secreted into portal blood by hypothalamic neurons, binds to receptors on lactotrophs, and inhibits both the synthesis and secretion of PRL.

Agents and drugs that interfere with dopamine secretion or receptor binding lead to changes in secretion of PRL. In addition to tonic inhibition by dopamine, PRL secretion is positively regulated by several hormones, including thyroid-releasing hormone (TRH), oxytocin, gonadotropin-releasing hormone (GrRH) and vasoactive intestinal polypeptide (VIP) [76, 77].

Moreover, estrogens provide a well-studied positive regulation of PRL synthesis and secretion [78, 79].

T

HE PROLACTIN RECEPTOR

The PRL receptor (PRLR) belongs to the class 1 cytokine receptor superfamily and they all share a homology in their extracellular regions, characterized by the conserved cysteine residues and the tryptophan-serine- x-tryptophan-serine motif [55]. The cytoplasmic domain of the PRLR lacks any intrinsic enzymatic activity; however, it includes a proline-rich motif (‘box 1’) that couples to protein kinase signaling molecules which in turn activate downstream effectors.

A single PRLR gene exists from which several PRLR isoforms derive. The PRLR isoforms differ in the length and composition and are referred to as long, intermediate or short PRLR with respect to their size. In human, one long, one intermediate and two short isoforms have been identified (reviewed in [55]). In rat, all three isoforms are present, whereas, in mice, one long and three short isoforms have been identified [80, 81]. Regardless of post-transcriptional splicing events, the extracellular ligand-binding domain is identical in all isoforms.

The PRLR binds to at least three types of ligands: PRL, PL, and primate GHs [57]. Activation of the cell surface receptor involves dimerization of two PRLR molecules [57], which is mediated by a single molecule of ligand [82]. The ligand binds in a two-step process in which site 1 on the PRL ligand molecule binds to one receptor molecule, after which a second receptor molecule binds to site 2 on the hormone, forming a homodimer

(18)

- 16 -

consisting of one molecule of PRL and two receptor molecules. [57]. Once bound to one of its ligand, PRLR triggers intracellular signaling cascades.

Like all cytokine receptors, PRLR lacks intrinsic enzymatic activity and therefore transduces its signal inside the cell via a wide number of associated kinases.

PRLR is virtually expressed in all tissues [55]. However, because of the extremely broad distribution of PRLR, it is currently difficult to propose a general overview of its regulation of expression [55].

P

ROLACTIN SIGNAL TRANSDUCTION

The main and best-known cascades involve the Jak/Stat pathway, the Ras- Raf-MAPK pathway, and the Src tyrosine kinases (e.g. Fyn), but other transducing proteins are also involved [55, 83]. Site-directed mutational studies have identified specific tyrosine residues within the PRLR cytoplasmic domain that can be phosphorylated and participate in recruiting Stats, insulin receptor substrates (IRS), and adaptor proteins to the receptor complex [55]. Depending on the presence or absence of these features, the various PRLR isoforms are expected to exhibit different signaling properties.

For example, the short PRLR is not tyrosine-phosphorylated, which prevents this isoform from interacting directly with SH2-containing proteins, such as Stat factors. However, these interactions may also be mediated by certain adaptor proteins [84]. The PRLR signaling pathways can be negatively regulated by protein tyrosine phosphatases, although their mechanism of action is still poorly understood [84, 85]. Recently, the SOCS (suppressor of cytokine signaling) gene family was identified and they function by negatively regulating the Jak/Stat pathway at the level of activation [86].

Finally, another emerging field in PRLR signaling is the occurrence of cross talk with members of other receptor families, such as tyrosine kinases [87, 88] or nuclear receptors [89].

ACTION OF PROLACTIN IN THE PROSTATE GLAND

PRL-mediated effects in the prostate are well described and supported by both in vivo and in vitro studies in rodent and human tissues. The presence of PRLR in both human and rodent prostate are well known [90-93]. Moreover, the PRL ligand has been demonstrated to be locally expressed both in human and rat prostate epithelium [93, 94]. The expression of PRL ligand in the rat DP and LP was found to be androgen dependent in vivo as well as in organ cultures [94]. These results could indicate a role for PRL as an

(19)

- 17 -

autocrine/paracrine growth factor, regulated by androgen, as well as mediating androgenic downstream effects in the rat prostate. Most of the described PRL prostatic effects have been studied in intact animals.

However, several reports indicate that PRL exert many androgen- independent effects [95, 96].

P

ROLIFERATION

In human BPH organ cultures, human primary prostate epithelium and in the androgen refractory human prostate cancer cell lines PC-3 and DU145, PRL has been shown to stimulate growth and significantly increase the cell proliferation rate [93, 97-99]. In one of these studies, DHT, estrogen and progesterone were assessed in parallel with PRL, but they all were found to exert weaker proliferating effects than PRL [99]. Moreover, PRL has been found to up-regulate ornithine decarboxylase (ODC) in the LP of rat. ODC is a rate-limiting enzyme in polyamine biosynthesis, and polyamines have been classified as growth mediators due to their effects on DNA- and RNA- synthesis in somatic cells [100, 101].

Several in vivo studies in rodents, have demonstrated the growth-promoting effects of PRL on the prostate [102-104]. To add to these studies are our own group’s generated PRL-transgenic mice, which develop a dramatic prostate enlargement [105, 106] (see the section RODENT MODELS OF PROSTATE DISEASE).

A

POPTOSIS

The concept of PRL regulation of target tissue size by controlling not only proliferative activity, but also apoptosis, is relatively new. PRL has been shown to significantly inhibit apoptosis in vitro in androgen deprived DP and LP prostate cultures, as assessed by nuclear morphology and in situ DNA fragmentation analysis [107]. This indicates a possible physiological role for PRL as a survival factor for prostate epithelium. In earlier in vivo work, a significant delay of castration-induced regression of the rat LP was noted in pituitary graft bearing animals [95, 108, 109]. In addition, these studies also indicated that AR did not mediate PRL actions on the prostate gland, as evidenced by the failure of flutamide, an AR antagonist, to inhibit the delay in prostatic regression. These results also reveal a lobe-specific response to PRL in the androgen-deprived prostate. Taken together, these observations suggest that in addition to known trophic actions in target tissues, PRL may regulate cell number by prolonging survival through anti-apoptotic mechanisms.

(20)

- 18 -

C

ITRATE PRODUCTION

The major function of the prostate gland is to accumulate and secrete extraordinarily high levels of citrate. In addition to citrate, the normal and BPH prostate also accumulate the highest levels of zinc in the body. In prostate cancer the capability for citrate production has been found to be lost and the ability for high zinc accumulation diminished [110, 111].

In several different species and model systems, PRL has been shown to androgen-independently stimulate citrate production, by direct regulation on enzymes involved in the citrate production, including mitochondrial aspartate aminotransferase, m-AAT, [112, 113], pyruvate dehydrogenase, PDH E1α [114, 115], m-aconitase [116] and aspartate transporter [14]. In addition, studies have revealed that the accumulation of zinc in the prostate also is regulated by PRL, independently of androgens. PRL increases both cellular and mitochondrial zinc levels of citrate-producing prostate cells [117]. Moreover, the regulation of the ZIP-type plasma membrane zinc uptake transporter has been reported to be regulated by PRL [118]. It is suggested that this ZIP-type zinc transporter is responsible for the ability to accumulate and transport high amounts of zinc in prostate cells.

PROLACTIN IN PROSTATE PATHOPHYSIOLOGY

Although PRL is well known to exert trophic effects on prostate cells, its role in the development and regulation of the age-dependent disorders, BPH and prostate cancer, is still poorly characterized. In order study the participation of PRL in the regulation of proliferative prostatic disorders several different experimental animal models have been used.

P

ROLACTIN IN HUMAN PROSTATE CANCER AND

BPH

The role of PRL in human prostate biology and pathophysiology is not well known. The altered endocrine status of aging men is likely to be of importance for development of prostate pathophysiology. Testosterone and GH levels decrease while estrogen levels increase with age. Conflicting data exists whether the circulating PRL levels increase or not with increasing age in the human male [32, 119-122]. Moreover, in a subset of aged men, an increase of TRH-stimulated PRL secretion together with an increase in circulating PRL level have been demonstrated [123, 124].

(21)

- 19 -

There is no clear correlation between serum PRL levels and risk of BPH or prostate cancer. More than 20 years ago, a beneficial effect of hypophysectomy in combination with castration compared to castration alone was observed in patients with disseminated prostate cancer [125, 126]. This indicated a role for one or more pituitary hormones, such as PRL, in advanced prostate cancer. Furthermore, significantly higher PRL serum levels have been reported in patients with prostate cancer [127, 128] and patients with BPH [129]. However, other studies report no differences in serum PRL levels between prostatic carcinoma patients and age-matched controls [121, 130].

Moreover, there is evidence that elevated PRL serum levels correlate with poorer prognosis in patients with advanced prostate [131, 132]. Although there are conflicting data, some clinical trials of advanced prostate cancer treatment have indicated a significant improvement in clinical response when combining conventional treatment with PRL suppression [128, 133-136].

There are clinical studies that have indicated increased prostatic tissue levels of PRL in patients with BPH [137] and prostate cancer [138]. Interestingly, PRL serum levels have been reported to transiently decrease following prostatectomy or transurethral resection of the prostate, TURP, [129, 139, 140], indicating loss of local PRL production or a prostatic influence on pituitary PRL secretion. Similar results have been presented in rodents [141].

Using immunohistochemistry, Nevalainen et al. reported local production of PRL in human prostate tissue [93]. Moreover, this study showed the presence of PRLR in the human prostate. The staining of the receptor was localized mainly to the secretory epithelium, but faint staining was also noted in the prostatic stroma. Collectively, these data provide significant support for the existence of an autocrine/paracrine loop of PRL in the human prostate. Furthermore, using in situ hybridization and immunohistochemistry Leav et al [91] demonstrated an increased PRLR expression levels in dysplastic lesions, whereas in lower grade carcinomas the receptor expression levels approximated those found in normal prostatic epithelium. Results from this study suggest that PRL plays a role in the development and maintenance of the human prostate and may participate in early neoplastic transformation of the gland.

E

XPERIMENTAL ANIMAL DATA

Enhanced growth of rodent prostate lobes after pituitary grafting under the renal capsule [102], or local grafting to a specific lobe [103, 104] has been reported. In rat, anterior pituitary grafting to the LP results in significant growth specifically in the LP compared to controls [103]. These results indicated a local direct effect of PRL on the LP. In mice, implantation of a

(22)

- 20 -

single anterior pituitary into the VP of intact mice results in a significant increase in VP weight and the area occupied by the glands of VP associated with the elevation of circulating PRL. Furthermore, hyperplastic lesions were noted in the grafted prostate lobes of these animals [104].

In another work, hyperprolactinemia has been reported to induce prostatic dysplasia in vivo. Noble rats, treated with testosterone and estradiol-17β2 for a prolonged time period, develop DLP dysplasia, a pre-neoplastic lesion. In these rats, the dysplasia was mediated via estradiol-induced hyperprolactinemia, as evidenced by effective inhibition of dysplastic evolution through the co-treatment of bromocriptin (a dopamine antagonist) [142].

Furthermore, animals which are exposed to a transient increase in PRL secretion prior to puberty have been shown to develop LP inflammation (prostatitis) as adults [143]. A recent study reports that early lactational exposure to atrazine, a toxic agent that suppresses suckling-induced PRL release, leads to altered PRL regulation and subsequent prostatitis in the male offspring. The mechanistic explanation is that without early lactational exposure to PRL (postnatal day 1-9), tuberoinfundibular neuronal growth is impaired and as a consequence prepubertal PRL levels become elevated.

This results in higher incidence and severity of LP inflammation in the offspring, evident at 120 days of age [144].

In addition to the abovementioned short PRL-treatment studies, also prolonged treatment of PRL has been shown to induce dramatic enlargement of the prostate as shown in our PRL transgenic mice which ubiquitously express the rat PRL transgene (Mt-PRL) [105] (see the following section).

RODENT MODELS OF PROSTATE DISEASE

Because the rodent prostate does not spontaneously develop prostate carcinoma and benign hypertrophy or hyperplasia, the usefulness of studying the mouse prostate as a model of human disease is frequently addressed.

However, the known heterogeneity of pathological prostate changes in the human prostate gland and the multifaceted nature of prostate disease have prompted the development of less complex, complementary model systems to study the etiology of prostate disease. Both prostate cancer and benign hypertrophy or hyperplasia can be induced in the rodent prostate through genetic modulation or chemical induction and several such models have been established. The advent of transgenic techniques in mice have put increasing

(23)

- 21 -

focus on the mouse as a model organism for in vivo studies aiming at understanding gene function and by this gain insights into human pathophysiological conditions. Moreover, the mouse genome project will soon be completed which will enable a direct comparison between the mouse and human genes.

T

RANSGENIC PROSTATE HYPERPLASIA MODELS

Male mice overexpressing the rat PRL gene, Mt-PRL transgenic mice, develop a dramatic enlargement of the prostate gland, which shows similarities prostatic hyperplasia in humans. These animals were generated using a construct consisting of the rat PRL gene under the control of the ubiquitous metallothionein (Mt) promoter, which gives the transgene a general transcription in virtually all cell types. Expression of transgene was detected in all parts of the prostate (DP, VP, LP, AP). The prostate enlargement is mainly characterized by an expansion of the stromal compartment and areas of glandular hyperplasia with accumulation of secretory material [105]. Although dysplastic epithelial features were detected in individual prostates from older PRL-transgenic animals, no development of prostate carcinoma has been observed. The PRL-transgenic animals display, in addition to high serum levels of PRL, approximately a three-fold increase in serum androgen levels compared to wildtype littermates. The degree of prostate enlargement showed no correlation to circulating levels of PRL or testosterone.

R

ODENT MODELS OF PROSTATE CANCER

There are several rodent models for human prostate cancer. One of the most well known is the Dunning-3327 rat prostatic adenocarcinoma model [145].

There are several recently established transgenic mouse models for use in prostate cancer studies [146]. The purpose of utilizing these animal models is to identify specific molecular changes in early malignant disease. As the mouse does not spontaneously develop prostate malignancy, different transgenic strategies for in vivo tumor induction have been developed including the use of the the SV40 early genes, such as the tumorigenic T antigen (Tag). Transgenes are usually under the control of a prostate-specific promoter region such as probasin or C3, directing expression to prostate epithelial cells.

The transgenic models of prostate cancer can be divided into two main types.

The first consists of models resulting from enforced expression of SV40

(24)

- 22 -

early genes. Two frequently used models are the TRAMP (transgenic mouse model for prostate carcinoma) model and the C3(1)-Tag transgenic model, which utilizes the minimal rat probasin promoter to drive the expression of the Tag gene. In addition, a number of transgenic lines use the long probasin promoter to express large SV40 early genes. These models are well characterized and widely distributed, displaying progressive disease ranging from epithelial hyperplasia or PIN to adenocarcinoma and development of metastases [147].

The second type of transgenic mice utilizes the promoters mentioned above to express various “natural” molecules that have previously been suggested to play a role in development of prostate cancer. The list is extensive but includes c-myc, Bcl-2 and dominant negative transforming growth factor beta (TGFß). Interestingly, the majority of these models only display a relatively mild phenotype, primarily epithelial hyperplasia or PIN. Moreover, these phenotypes usually not arise until the mice are of advanced age.

O

THER GENETICALLY ENGINEERED MOUSE MODELS WITH PROSTATE PHENOTYPE

Mouse models genetically engineered in the prolactin signaling pathway Null mutated mice have been generated both for the PRL ligand, PRL-/-[148], and the PRLR-/- [149]. PRL-/- males are reported fertile [148], whereas studies of male PRLR-/- mice have demonstrated both a subset of completely infertile males and a general latency to first successful mating [150].

Moreover, the studies of the prostate gland in PRLR-/- males did reveal only subtle histological alterations and the PRL-/- prostate has not been very well characterized. Taken together, the data from these two knockout mouse models indicate that PRL action is not of essential importance for male fertility and normal anatomical development of the prostate gland. However, studies of more functional aspects of the gland need to be carried out in these animals.

PRL can activate several of the Stat proteins, including Stat 1, 3, 5a, and 5b, but the two latter acts as the major mediator [55]. Stat5a-/- and Stat5b-/- knockout mice have confirmed these molecules as the major transducers of PRL signaling in both prostate and mammary gland [151], and also shown similar phenotype to those of the PRL-/- and PRLR-/- knockout mouse models, mainly emphasizing the irreplaceable role of PRL in reproduction and mammary gland development. PRL signaling in rat prostate tissue is primarily transduced via Stat5a and Stat5b, likely supporting the viability of

(25)

- 23 -

prostate epithelial cells during long-term androgen deprivation [152]. In the prostate, studies in Stat5a-/- knockout mice have provided evidence for a direct role of Stat5a in the maintenance of normal tissue architecture and function of the mouse prostate [153]. Lack of Stat5a function results in a distinct prostatic phenotype characterized by an increased occurrence of cyst formation with disorganization and detachment of prostate epithelial cells. In addition to PRL, other polypeptide factors, such as GH, insulin-like growth factor I (IGF-I), epidermal growth factor (EGF) and interleukin-6 (IL-6) are known to activate Stat5.

Mouse models genetically engineered in other hormones

The AR transgenic mice overexpress the AR specifically in prostate secretory epithelium [154]. The earliest alteration observed in the AR transgenic mouse prostates was an extensive 5-fold increase in the proliferation of secretory epithelial cells, as evidenced by immunostaining of the proliferating marker Ki-67, in the absence of histological abnormalities.

Proliferation in these glands was associated with increased apoptosis, possibly accounting for the absence of hyperplasia. Older AR transgenic mice developed focal areas of intraepithelial neoplasia, resembling human high-grade PIN, but no further malignancy has been observed. A certain resistance to malignant transformation in the mouse prostate compared to humans has been suggested. No reports of any tumorigenic effects of exogenously added androgens in these models are available.

The recent generation and characterization of the various estrogen modulated mouse models (αERKO, βERKO, αβERKO and ArKO) have provided new insights regarding the role of estrogens in prostate growth and development [155]. A specific direct response to estrogens is the induction of changes in the prostatic epithelium, termed squamous metaplasia [156-159]. Tissue recombinant studies using epithelium and stroma from wildtype and transgenic mice lacking a functional ERα (αERKO) or ERβ (βERKO) have demonstrated that the development of squamous metaplasia is mediated through stromal ERα [160, 161]. Furthermore, a distinct phenotype of focal epithelial hyperplasia in the VP has been reported in aging mice lacking functional ERβ (βERKO) [162, 163], while no apparent prostate pathology or enlargement has yet been reported in αERKO or the double knockout αβERKO [155]. Altogether, these findings indicate an anti-proliferative role for epithelial ERβ and also suggest that an unbalanced stromal ERα in action could contribute to the phenotype observed.

The ArKO (aromatase knockout) mouse lacks endogenous estrogen

(26)

- 24 -

production due to a non-functional aromatase enzyme. In the ArKO mouse, the combined effects of estrogen absence and elevated androgen and PRL levels result in a moderate prostate enlargement with hyperplasia evident in all lobes and tissue compartments [161]. Moreover, an associated up- regulation of epithelial AR was demonstrated in the ArKO mouse and has been suggested to contribute to the observed phenotype. In the absence of endogenous estrogen (ArKO) or ERs (αERKO and βERKO), prostate development occurs normally, suggesting that intact estrogen signaling is not essential for the initiation of neonatal prostate growth. The histological appearance of the prostate hyperplasia in ArKO male mice is strikingly similar to that of the Mt-PRL-transgenic mice.

In contrast, the AROM+ mice, which overexpress the aromatase gene, resulting in elevated estrogens levels, combined with significantly reduced testosterone and FSH levels, and elevated levels of PRL and corticosterone [164]. AROM+ males present a multitude of severe structural and functional alterations in the reproductive organs. Furthermore, squamous metaplasia has been seen in the prostatic collecting ducts, consistent with high levels of endogenous estrogens. Some of the abnormalities, such as non-descended testes and undeveloped prostate, resemble those observed in animals exposed perinatally to high levels of exogenous estrogen, indicating that the elevated aromatase activity results in excessive estrogen exposure during early phases of development.

HORMONE/GROWTH FACTOR REGULATION OF THE PROSTATE

All lobes are responsive to both estrogens and to androgens, but to varying degrees; the VP is more sensitive to androgens and the AP more sensitive to estrogens [159, 165]. In rat prostate, both testosterone and estrogen have been shown to regulate the level of the long PRL receptor mRNAs in a tissue- specific manner [92]. In addition to steroid hormones, several different growth factors and other pituitary hormones have been shown to regulate cellular growth, differentiation and apoptosis.

A

CTION OF ANDROGENS IN THE PROSTATE

Androgen is a critical factor for the survival of prostatic epithelial cells.

Underdeveloped prostate gland is seen in eunuchs who lack androgen stimulation since childhood [166]. Castration-induced androgen-withdrawal

(27)

- 25 -

regress the the number of epithelial cells in the prostate gland via an active process of apoptosis [167, 168]. Apoptosis can be observed within one day after castration and nearly 2/3 of epithelial cells are lost in the VP by seven days of castration [169]. In contrast, testosterone replacement to castrated rats stimulates the re-growth of the gland to its normal size via proliferation of new epithelial cells from basal cells [170].

I

NTERACTIONS BETWEEN PROLACTIN AND ANDROGENS IN THE PROSTATE GLAND

PRL has been shown to potentiate the action of androgens in the support and stimulation of prostatic growth and metabolism [171-173]. This has been hypothezised to be accomplished through increasing prostate receptivity to androgens, mainly by affecting AR levels and 5-alpha reductase activity.

Results suggest that PRL is involved in regulating AR synthesis, at least partially by direct action on the prostate gland. In immature, hypophysectomized male rats, PRL treatment can significantly increase AR mRNA levels [174]. Findings in adult, castrated and pituitary grafted rats suggest that PRL promotes LP growth via an increase in nuclear AR levels, and thus optimizes tissue response to circulating testosterone [175].

Furthermore, pituitary grafting in immature rats can produce a significant increase in the weight of the seminal vesicles and the VP and AP [176]. In the VP, nuclear AR content increased, whereas the cytosolic AR content decreased, suggesting increased translocation of the AR to the nucleus. In a study on human BPH patients, cytosolic and nuclear levels of AR were shown to be proportional to plasma PRL levels [177]. These findings indicate plasma PRL involvement in the regulation of AR content also in the benign human prostate.

Recently the existence of crosstalk between the signal transduction systems of steroid hormones and peptide hormones/growth factors were recognized [178-180] which provides a mechanism for locally produced growth factor influence on AR activation. In the progression of prostate cancer to an androgen-independent state, local growth factors, such as PRL, may prove instrumental in regulation of cell growth.

In rat, hyperprolactinemia by pituitary grafting can lead to increased 5-alpha reductase activity in the testis [181] but indications of a PRL-induced increase in 5-alpha reductase activity in the prostate is limited [182]. PRL mediation of steroid uptake through alterations of the plasma membrane permeability in human BPH tissue has also been reported [183].

(28)

- 26 -

To interpret findings in rodent versus human studies, one needs to be aware of the important differences in influence of PRL on circulating androgen levels.

In man, PRL is known to decrease circulating androgen levels through depression of gonadotrophine release from the pituitary gland [184], whereas in rodents, PRL can elevate circulating androgen levels by increasing the response to luteinizing hormone in the testis [185].

A

CTION OF ESTROGENS IN THE PROSTATE

A hierarchy of estrogen responsiveness in the three prostatic lobes has been revealed in male mice, with the AP being the most responsive, the dorsolateral lobe less responsive, and the ventral lobe the least responsive.

[159].

The expression of both known estrogen receptor subtypes in adult human and rodent prostate is now well established, with expression of ERα described primarily in a subset of stromal cells and ERβ restricted to the ductal epithelium [186-188]. Although the newly discovered ERβ shares many of the functional characteristics of ERα, the molecular mechanisms regulating the transcriptional activity of ERβ may be distinct from those of ERα. For example, the growth effects of estrogens during fetal development are mediated primarily by ERβ in the human prostate, which can be immunodetected in the nuclei of nearly 100% of epithelial and in the majority of stromal cells throughout gestation. However, ERα has been shown to contributes to postnatal glandular development [156].

Estrogen plays an important role both in prostate physiology and pathophysiology. The developing prostate is particularly sensitive to estrogenic exposure. During prostate morphogenesis, elevated levels of endogenous (maternal or excess local production) or exogenous (diethylstilbestrol or environmental chemicals) estrogens induce permanent changes in prostate growth in rodents. Fetal and neonatal exposure to estrogens results in pathological and functional changes of the prostate [189].

High-dose of testosterone together with estradiol stimulates prostatic carcinogenesis in adult male rats [190]. In mice, these effects are dose- related as low-dose estrogen exposure may increase the adult prostate size whereas high-dose exposure reduces prostate size [189]. An increase in AR levels has been associated with low-dose estrogen-induced increases in prostate size [190]. Neonatal exposure of rodents to high doses of estrogen is known to permanently imprint the growth and function of the prostate and predispose the gland to hyperplasia and severe dysplasia analogous to PIN with aging [160]. Following neonatal exposure of rats to high doses of

(29)

- 27 -

estrogen on days 1-5 of life, a permanent reduction in prostate growth and responsiveness to androgen occurs relative to a reduction in AR expression in adult animals [165]. Moreover, exogenous estrogen administration in adult rodents leads to squamous metaplasia of the AP [157, 159]. As mentioned earlier, development of squamous metaplasia has been shown to be mediated through stromal ERα [160, 161].

I

NTERACTIONS BETWEEN PROLACTIN AND ESTROGENS IN THE PROSTATE GLAND

Estrogen are known to act directly on pituitary lactotrophs and indirectly on the hypothalamic dopaminergic system and several studies suggest that neonatal estrogen treatment can induce long-term alterations in pituitary synthesis and release of PRL [191-193]. Moreover, estrogens are well- known to promote PRL release resulting in elevated PRL levels systemically [78, 79]. It is thus quite possible that the prostate effects of estrogen imprinting are in fact partially PRL-mediated. Furthermore, PRL is able to stimulate expression of both ERα and ERβ in corpus luteum and decidua during pregnancy [194-196] as well as stimulate estradiol binding activity or mRNA levels in the mammary gland [197] and liver [198]. In the prostate, effects of estrogen treatment appear to be in part mediated by increased PRL levels [199], something that is further demonstrated in the aforementioned dysplastic prostate model of estrogen-treated Noble rats [142].

A

CTION OF OTHER PEPTIDE HORMONES AND GROWTH FACTORS IN THE PROSTATE

Growth factors regulate cellular growth, differentiation and apoptosis. In addition to steroid hormones, an array of positive and negative growth factors controls the balance between cell proliferation and apoptosis in the prostate. Several oncogene products that contribute to neoplastic proliferation have been found to be homologues to growth factors, growth factor receptors, or molecules in the signal-transducing pathways of these receptors. There are numerous growth factor families that have been implicated in normal, neoplastic and malignant prostate growth and it is far beyond this thesis to review the action of all reported hormones and growth factors. The in the literature mentioned growth factors include, the IGF family, EGF, TGF, FGF family, platelet-derived growth factor (PDGF) and VEGF, which all are the main stimulatory regulators of proliferation in the prostate [200]. Furthermore, the pituitary hormones, GH and luteinizing

(30)

- 28 -

hormone (LH), play physiologically significant roles in the normal prostate, either alone or synergistically with androgens [201]. Nevertheless, the involvement of these hormones in the development of BPH and prostatic carcinoma is an issue that needs to be addressed.

The TGF-family is the main inhibitory regulator of proliferation acting on the epithelial cells. However, recent studies have demonstrated proliferative and anti-apoptotic effects of TGFβ in stromal cells [202].

Altogether, the growth factors exert autocrine and paracrine effects upon stromal and epithelial cells and interact with other factors and binding proteins to control prostate growth [203].

FUNCTIONAL GENOMICS IN THE STUDY OF THE PROSTATE GLAND

The network of action of different hormones and growth factors on the prostate gland and their involvement in prostate pathophysiology are unquestionable complex. The recent completion of the human [204], and the draft of the mouse [205], genome sequence together with the improvement of high-throughput technologies, such as gene expression profiling, will hopefully provide a basis for rational determination of which pathways and molecular targets that are appropriate to further study. The unveiling of a detailed genetic map of the main species and models of prostate research promise to dramatically increase our understanding in the genetic basis of prostate disorders together with the basic mechanism of the action and involvement of hormones and growth factors for the induction of prostate disease.

(31)

- 29 -

AIMS OF THE THESIS

The overall aim of this thesis was to study the consequences of chronic exposure to extraordinary high levels of PRL in the development of prostate hyperplasia as well as to characterize the molecular mechanisms present in the hyperplastic prostates.

The specific aims were:

• To investigate the role of circulating androgen in the promotion of prostate hyperplasia in PRL transgenic mice with transgene onset during early prostate development (Paper I)

• To characterize a new PRL-transgenic model of prostate hyperplasia in which the PRL transgene was overexpressed specifically in the prostate with onset at puberty (Paper III)

• To compare the ductal development in two models of prostate hyperplasia; one with fetal onset and the other with pubertal onset of the PRL transgene expression (Paper III)

• To evaluate the use of differential gene expression analysis in characterization of the molecular mechanisms of the prostate hyperplasia in PRL transgenic mice (Paper II and IV)

(32)

- 30 -

METHODOLOGICAL CONSIDERATIONS

TRANSGENIC ANIMALS

The establishment of the transgenic technology has introduced new and invaluable techniques to study and understand the function of a specific gene in biological processes. There are basically two types of transgenic animals based on the technique used to generate them. The first method to be established in 1980 was the microinjection technique allowing overexpressing of a gene product by injection of foreign DNA into a one cell mouse zygote [206]. The incorporation of the foreign DNA is completely random using this approach and it is only possible to overexpress a gene product and not to mutate a certain gene. In contrast, the other embryonic stem cell (ES)-cell technique (also known as gene knockout) was established in 1987 and this method made it possible to interact with the mouse genome at a specific position and to mutate a specific gene [207]. A wide range of transgenic and knockout mouse models have now been established and further technical improvements have made both temporal and spatial overexpression/gene deletion possible. These accomplishments have given unique insights into the specific biological properties and functions of specific genes and furthermore provided valuable models for investigating the functional in vivo role of target genes.

In this thesis we utilized two different transgenic mice models. The rat PRL transgene where used in both constructs, in parallel with two different promoters to direct spatial (where) and temporal (when) expression of the transgene, resulting in two different PRL transgenic mouse models. In paper I and II, the metallothionein (Mt) promoter was used to drive the PRL transgene. The Mt gene is expressed in virtually all cell types. Activation of the Mt-1 promoter during the early embryonic stage is well described, with abundant expression already by day 12 of gestation reported [208, 209].

Thus, the PRL expression was considered general in the Mt-PRL transgenic mice. In contrast to the general expression of a transgene, a cell-specific promoter can be used that direct the expression of the transgene to a certain cell type. In paper III and IV, the probasin (Pb)-PRL transgenic mice were utilized. The construct of these mice include the minimal Pb promoter to direct the expression of the rPRL transgene to the epithelial cells of the DP, LP, and VP [210]. Pb is an androgen-dependent basic secretory protein, abundantly localized in the lumen and acinal regions of the rat prostate epithelium [211]. Studies have demonstrated that the Pb minimal promoter (458 bp) can target heterologous gene expression specifically to the prostate

Molecular Characterization of Prostate Hyperplasia in Prolactin Transgenic Mice