As described in the previous section the prostate consists of two major cellular types, the stroma and the epithelium. The stromal component is comprised of smooth muscle cells, fibroblasts and endothelial cells; 5 cell sub-types, basal epithelium, secretory epithelium, transit amplifying cells, neuroendocrine cells and stem cells, constitute the epithelium Prostate cancer is non-regulated proliferation primarily of the epithelium. A number of factors may contribute to dedifferentiation and proliferation of epithelial cells, including aberrations in surrounding stromal cells via cytokines and growth factors secreted by both stromal and epithelial cell types (18,19).
To date, the exact triggers for development of prostatic adenocarcinoma have not been elucidated, although epidemiological studies have shown links to both environmental (discussed subsequently) and genetic contributors. A variety of growth factors and cytokines have been shown to influence the growth rate and development of prostatic cancers, and a number of proteins and genes have been identified whose regulation and/or function are altered between malignant and benign states. Several of these proteins and genes have been flagged as potential biomarkers and therapeutic targets for prostate cancer.
This chapter will provide an overview of the current literature, and will discuss factors that contribute to the development and progression of PCa, as well as the potential of new biomarkers for diagnosing and treating this disorder.
Genetic aberrations, whether inherited or acquired, are universally associated with cancer development, growth, and progression. Several cancer types, such as breast, renal and colorectal, have been shown to have a familial component (20). In an effort to identify important genetic determinants associated with hereditary prostate cancer, a number of research groups have employed different methodologies involving large-scale populations (epidemiology), families (with one or more affected individuals) and twins.
Major genetic epidemiologic studies published in the last two decades support the notion that prostate cancer may exist as clusters in families. In the 1980s, a Utah Mormon genealogy study found that prostate cancer exhibited the fourth strongest degree of familial clustering after lip, melanoma, and ovarian cancers . Prostate cancer, interestingly, had a higher familial association than either colon or breast carcinoma, which are known to be predisposed by genetic or familial components. A later study, determined cancer pedigrees in 691 men with prostate cancer and 640 spouse controls, and found a positive familial history of prostate cancer . They concluded that men with an affected father or brother were twice as likely to develop prostate cancer as men with no affected relatives. Although these studies strongly suggest that familial clustering of prostate cancer risk does exist, they did not address the underlying aetiological mechanisms. Indeed, familial clustering can reflect either shared environmental and lifestyle risk factors, or a genetic mechanism, or indeed both.
In order to address the inherent difficulties in separating the inheritable and environmental causes of prostate cancer, the International Consortium for Prostate Cancer Genetics (ICPCG) was formed. In early 2005 they published a world-wide study of 1,233 families located in North America, Australia, and Europe. While the study included some families with Asian, Hispanic, Native American and African-American backgrounds, 1,166 of the 1,233 families were Caucasian. The researchers found 5 regions (5q12, 8p21, 15q11, 17q21 and 22q12) in human chromosomes that potentially could harbour cancer susceptibility genes. In addition, the authors flagged chromosomal regions which linked to families with 5 or more affected members (22q12, 1q25, 8q13, 13q14, 16p13 and 17q21) and those which corresponded with early onset of disease (less than 65 years; 3p24, 5q35, 11q22, and Xq12) . While, to date, there are no known cancer susceptibility genes in these loci, work is ongoing to refine the mapping of these regions to the gene level.
Twin studies compare the similarities (concordance rate) of a trait or traits, traced to either monozygotic (MZ) or dizygotic (DZ) twins to dissect the relative genetic and environmental influences of a disease. Several twin studies (although with limited subject numbers) of prostate cancer reported higher concordance rates in MZ twins compared to DZ twins, implicating a genetic contribution for prostate cancer. Lichtenstein et al (2000) (24) analyzed Swedish, Finnish and Danish twin registries and concluded that the recurrent risk in the twin of an affected man was 21.1% for MZ and 6.4% for DZ. More recent mathematical modeling of genetic risk of a number of cancers, by the same research group (25), found that the contribution of genetic susceptibility to prostate, breast, and colorectal cancers was small to moderate. Although these aforementioned studies clearly demonstrated a potential genetic basis of prostate cancer, albeit low, for a portion of familial prostate cancers, this rate is much lower than coeliac disease where strongly heritable components were found in 75% of the MZ with 11% prevalence in first-degree relatives . The significant differences in the contribution by the heritable element on the development of prostate cancer, as opposed to coeliac disease, suggest the importance of gene-environment interaction, which may be pivotal in prostate carcinogenesis.
In addition to hereditary prostate cancer-associated genes, a number of studies have focused on genetic components that may be associated with prostate cancer development. Association studies are generally case-control studies based on the comparison of frequencies of an allele in unrelated cases and normal controls. A significant difference in allele frequency between cases and controls can be expected if a particular allele harbours causal variations or if that allele is closely linked with (or in close proximity to) a disease-causing mutation. These types of association studies are susceptible to aetiological (genetic versus environmental), inheritance (dominant, recessive, or X-linked), and locus heterogeneity, as well as allelic (same gene but different mutations) and founder (same mutation in different genetic background) heterogeneity. Despite these difficulties, association studies have identified genes that may be associated with, or contribute to, prostate cancer development and progression.
BRCA1 and BRCA2 genes, linked to breast cancer development, have also been examined in prostate cancer due to links between these cancers’ aetiology, observed through epidemiologic, biologic, and molecular studies (27, 28). Recent association studies demonstrated an increase in the frequency of BRCA1 and BRCA2 mutations in prostate cancer carriers . The most recent studies have shown that the rate of prostate cancer diagnosis after age 60 in cases with a deleterious mutation in BRCA1 or BRCA2 was 5.2% compared with 1.9% of controls. . In addition, data reported thus far show that mutations of BRCA1 and BRCA2 may increase the risk of prostate cancer, particularly in the case of BRCA2 for early-onset disease. The contribution of germ line mutations in these genes to familial clusters of prostate cancer needs to be more fully assessed.
Androgens serve as a most important ligand for prostate cancer, stimulating growth via the androgen receptor action pathway. The androgen receptor action pathway, broadly defined, includes the androgen receptor, the 5-alpha-reductase gene and the various members of the cytochrome p450 family.
Mutations within the AR gene are detected in 10 – 20% of prostate cancer specimens, with mutations more common in hormone escape than hormone sensitive specimens . The androgen receptor (AR) is comprised of three important domains: N-terminal transactivation domain, DNA binding domain, and C-terminal steroid binding domain. Mutations in the N-terminal transactivation domain include variations in the number of polymorphic triplet repeats located within this area of the gene . Several studies have suggested that AR genes with shorter repeat lengths may increase the risk of developing more aggressive prostate cancer, as the number of polyglycine repeats is inversely related to the ability of the AR to activate target genes . Mutations at the steroid-binding domain have been associated the promiscuous nature of AR which recognizes non-androgen ligands, such as oestrogen, progesterone, and anti-androgens as agonists that confer AR-induced gene transactivation and growth stimulatory activities (36). Overall, the increase in incidence of mutations and the variety of mutations observed in the AR in metastatic cancers, suggest that these regions of AR heterogeneity may be acquired through gene-environment interactions.
Amplification or duplication of the AR gene has been associated with the transition from hormone sensitive to hormone refractory tumours. Edwards et al, (2003) (37) found that less than 5% of hormone sensitive prostate cancers contained more than one copy of the AR gene, compared with 20-30% of hormone resistant tumours. In 80% of these cancers there was a corresponding increase in AR protein concentrations. However, it was the presence of the gene duplication and not just higher AR protein levels which significantly predicted poorer survival rates for patients in their study .
Alterations in the 3-beta-hydroxysteriod dehydrogenase gene (also a member of the AR pathway) have also been strongly associated with hereditary prostate cancer . 3-beta-hydroxysteroid dehydrogenase metabolises dihydrotestosterone to 3-alpha- and 3-beta-diol. Recently Guerini et al (2005) have found that 3-beta-diol does not bind androgen receptors but efficiently binds the estrogen receptor (ER-beta), and exerts an inhibitory effect on prostate cancer cells through the activation of ER-beta signaling. Subsequently, deletiorous mutations in the 3-beta-hydroxysteriod dehydrogenase gene are thought to confer a growth advantage to cancer cells.
Although the androgen receptor pathway dominates in hormonal-tumour interactions in prostate cancer, other relationships are also operative. As well as oestrogens having an increasingly recognised role, the growth hormone secretagogue ghrelin and a preproghrelin isoform have been shown to be highly expressed in prostate cancer with activation via the protein kinase pathway (39). Both inhibin and activin subunits are known to be expressed in the normal prostate with the inhibin alpha subunit gene down regulated in prostate cancer with a loss of heterozygosity at the gene locus and methylation of the promoter (40). Overexpression of the activin betaC-subunit has been demonstrated to increase activin AC heterodimer levels, concomitantly reducing activin A levels and decreasing activin signaling in PC3 cells (41).
To determine what differences might distinguish hereditary prostate cancer from its sporadic counterparts, a number of clinical features of prostate cancer were examined by Carter, et al. . Clinical stage at presentation, pre-operative PSA, final pathologic stage, and prostate weight were examined in a series of approximately 650 patients divided among three categories. Individuals were classified as having hereditary disease if 3 or more relatives were affected in a single generation, prostate cancer occurred in each of 3 successive generations in either paternal or maternal lineages, or 2 relatives were affected under the age of 65 years. For the other groups, either no other family members were affected (sporadic disease), or other family members were affected but not to the extent found in families classified as hereditary. In summary, no unique clinical or pathological characteristics distinguished hereditary prostate cancer in this group of patients. This parallel between hereditary and sporadic prostate cancer extends to the incidence of multifocality found in both of these categories.
While mutations or gene duplications have been associated with familial prostate cancers, other molecular changes have been documented in the majority of prostate tumours. Although the mechanisms behind these changes have not always been elucidated, it is believed that changes in gene methylation patterns, gene expression profiles and expression of non-coding mRNA’s play important roles in cancer development and progression.
One of the mechanisms used by cells to alter gene expression is by conjugation of a methyl group on certain cytosine residues in the gene promoter. In a normal cell this may function to maintain cellular differentiation, but recent reports have described several genes which are abnormally silenced, or activated, in many tumour types including prostate cancer, by promoter methylation. . Lodygin et al, (2005) (43) described many genes, including cell cycle regulating genes, angiogenesis inhibitors, apoptosis inducing genes, growth factor receptors and transcriptional control genes that are “switched off” in prostate cancer. The “switch” could be thrown into reverse by treatment with 5-aza-2’ deoxycytidine, a chemical that removes methyl groups from cytosine residues. A study of 41 primary tumours from prostate cancer patients found that frequencies of CpG methylation detected in the promoter region of a selection of the above genes were as follows, GPX3 (Glutathione peroxidase – detoxification), 93%; SRFP1 (Secreted frizzled-related protein 1 – signaling), 83%; COX2 (cyclooxygenase 2 –signaling) 78%; DKK3 (Dickkopf homologue 3 – signaling) 68%; GSTM1(Glutathione S-transferase M1 - detoxification), 58% and KIP2/p57 (cell cycle control), 56%.
The authors also found that methylation in some gene promoters did not always equate to reduced transcription rates, and that some genes apparently “switched on” after 5-aza-2’ deoxycytidine treatment did not show detectable CpG methylation. They concluded that silencing of genes by CpG methylation occurs at an early stage of prostate cancer development. This may have implications for the use of certain methylation events as diagnostic markers for prostatic disease (44).
Recently, global methylation in the chromosomes of prostate cancer cells has been studied with specific antibodies that detect all CpG sites, both in gene promoters and elsewhere in the chromosomes (45). Interestingly, the overall amount of methylation in tumour cells was decreased when compared with benign cells from the same individuals. It was also noted that men with recurrent disease and/or positive surgical margins showed overall higher levels of methylation in their tumour cells compared with men with early stage, surgically curable disease (45). Brothman et al, (2005) (45) suggest that genomic hypermethlyation is associated with condensed DNA. Hypomethylation of these regions permits a reduction of DNA condensation and may result in chromosomal rearrangement and genetic instability. In addition they suggest that global DNA hypomethlyation precedes hypermethylation of specific gene promoters (such as GSTP1) in cancer development, and the balance between the two methylation states may be critical in determination of aggressive vs non-aggressive disease (45). However, whether CpG methylation is a cause or consequence of cancer remains to be shown.
With the rise of microarray technologies, analyses of changes in multiple gene transcription rates have been described. Interestingly, some changes in the transcriptional output of the human genome points to the existence of a significant number of non-coding RNA transcripts, derived from intronic genomic regions, with some of these being oriented antisense relative to the protein-coding mRNA of the gene. To investigate the expression level of intronic messages in human tissues, a subset of approximately 2,000 totally intronic Expressed Sequence Tags (EST) clusters and 2,000 clusters from exonic segments of known genes, was selected for intronic microarray analysis. Hybridization of these intronic microarrays with 27 prostate tumours and corresponding adjacent normal tissue revealed that in prostate cancer, the fraction of expressed messages arising from exonic or intronic transcripts were similar . Moreover, the expression levels of 23 intronic non-coding transcripts correlated (p value - 0.001) with the degree of prostate tumour differentiation. It has not been determined whether the expression of intronic antisense RNAs in tumours is a true mechanism of cancer disregulation or just reflects broad errors in promoter recognition/transcription initiation. However, regulatory elements have been identified for a number of non-coding RNA’s, including DD3/PCA3 and PCGEM1 – see below.
While a variety of genetic and environmental determinants may contribute to the cause of malignancy, cancer cells display similar phenotypes and modes of development such as changes in cell adhesion, attachment, migration and invasion. These phenotypes and the genes controlling them are in turn regulated by epigenetic factors, such as soluble proteins, steroids and growth factors released by immune cells and adjacent stromal cells (e.g. fibroblasts, smooth muscle cells, endothelial cells, and neuroendocrine cells) and alterations in the extracellular matrix (ECM) surrounding tumor epithelium .
Reciprocal cellular interaction between stroma and epithelium is involved in foetal prostate development, post-natal prostate growth and maturation, maintenance of differentiation status, hormonal responsiveness, and the aging and senescence of prostate gland in adulthood ). During neoplastic progression, key phenotypic changes in prostate cancer cells are also modulated by a dynamic, two-way communication between tumour epithelium and various stromal cells, including fibroblasts, smooth muscle cells, vascular endothelium and bone-derived cells including osteoblasts . All of these cell types and the factors they secrete form the cell’s microenvironment that can either enhance or repress cancer development and progression. One of the challenges in understanding carcinogenesis is determining how the tumour microenviroment interacts with genetic changes in the cells themselves.
Experimental evidence using co-cultured human benign prostatic hyperplasia (BPH)-derived stromal and epithelial cells indicate that the expression of AR, PSA, and 5 alpha-reductase in epithelial cells relies on the inductive influence of neighbouring stromal cells . Similarly, in co-cultured rat prostatic epithelial and fibroblast cells, the androgen responsiveness of prostate epithelial cells can be conferred by the presence of fibroblastic cells . Therefore the action of AR in prostate development is mediated by the local stroma. In addition, oestrogens have been shown to have a role in cancer development, yet the oestrogen receptor is predominantly localized in stromal cells. . Animal models have demonstrated prostatic cell proliferation occurs in androgen deficient mice due to exogenous oestrogen. . In addition, elevated androgen levels in oestrogen deficient mice induced prostatic hyperplasia . While neither of these hormones alone induced prostate cancer, combined androgen and oestrogen therapy did evoke prostatic dysplasia and adenocarcinoma. It is proposed, therefore, that both androgens and oestrogens influence the process of prostate carcinogenesis .
The function of androgen ablation therapy (ABT) is to prevent activation of AR regulated genes. Failure of ABT, as in hormonally resistant prostate cancer, was logically linked to genetic modification of the AR, as described above. More recently however, there has been recognition of the effect of post-translational modifications on the AR . In 1998, Blok et al. demonstrated phosphylation of the AR in response to binding by the ligand DHT (dihydrotestosterone). In addition to stabilizing the active AR homodimers, this phosphorylation can influence AR mediated gene activation . In vivo, phosphorylation of AR by MAPK (serine-threonine kinase) and AKT (protein kinase B) sensitizes AR to low levels of DHT, allowing low levels of androgens, or alternative steroids to induce translocation of the AR to the nucleus, facilitating gene activation . MAPK has been shown to be activated in cell lines derived from hormone-refractory tumours, and is correlated with advanced stage and grade in prostate cancer . While the exact mechanism is not fully known, transfection of cells with c-Ras (cell growth regulator; activated in many cancer types), leads to increased activation of MAPK and development of hormone escape . AKT can be activated via PI3 kinase pathway (involved in cell growth, adhesion and migration in many tumours; (69), and specifically phosphorylates AR at Ser210 and Ser790 (70)– see Figure 1
The control of AR function also involves interaction of the receptor with a number of co-factors or proteins that bind either as part of a complex or directly to AR and increase (co-activators) or inhibit (co-repressors) the transcriptional activity of the AR.
Co-activators include CBP (CREB-binding protein), beta-catenin, ARA55 and ARA70 which all act to alter ligand specificity of the AR . These mechanisms include allowing the antiandrogens such as hydroxyflutamide (CBP, ARA70) and bicalutamide (ARA70) to act as agonists and/or by permitting low concentrations of adrenal androgens or oestradial to activate AR (ARA55, beta-catenin, ARA70; . Some co-activators activate AR in the absence of ligands, such as SRC-1, SRC-3, p300, Tip60, and c-Jun. C-Jun binds to the AR, promoting homodimerization and subsequently activating AR dependent transcription initiation of downstream genes (such as PSA) . SRC-1 and SRC-3 are members of the steroid receptor cofactor family, which is commonly overexpressed in hormone-refractory prostate cancer. This family of proteins normally facilitates AR transcriptional activity in the presence of androgens, however, phosphorylation of SRC-1 by MAPK (activated as discussed above), may be one mechanism by which SRC-1 activates AR in the absence of androgens. Tip60 is linked to the transcription activation activity of AR by inducing changes to AR acetylation. In cell line models, in the absence of androgens, p300 is required for Il-6 (see below) stimulated growth, and it has been proposed that p300 plays an important role in the development of hormone refractory tumours .
Figure 1. Phosphorylation of the androgen receptor (AR) by MAPK and AKT. Numerous signalling factors (intracellular and extracellular), involved in inducing cell growth and proliferation, stimulate the Ras pathway to activate MAPK. Subsequently, MAPK phosphorylates AR, enabling AR to form dimers, enhancing ARE (androgen response element) dependent gene expression. Similarly, intra- and extra-cellular factors inducing cell growth and migration and inhibiting cell adhesion activate the PI3 kinase pathway resulting in AKT dependent phosphorylation, dimerisation and activation of AR.
A less well understood mechanism of AR co-activators involves movement of AR to the nucleus. In LNCaP cells, STAT3 binds ligand-free AR and facilitates its translocation to the nucleus, and it is via this mechanism that the STAT3/AR complex activates AR dependent genes (in the absence of androgens) in response to IL-6 stimulation .
The mediation of stromal-epithelial interactions in the normal and malignant prostatic environment involves a number of soluble factors that can serve paracrine, autocrine or intracrine functions (72). Several soluble factors have been identified performing a variety of functions from angiogenesis, growth enhancement, and dedifferentiation. These are listed in Table 1 (72).
Table 1. Most commonly cited soluble factor signaling pathways regulating prostate growth and differentiation (from Chung et al., 2005 [19])
|
Soluble Growth Factor |
References |
|||||
|---|---|---|---|---|---|---|
|
Name |
Source |
Receptor |
Receptor Location |
Function |
Regulation at Androgen Independent Progression |
|
|
VEGF |
Epithelium, stroma |
VEGFR-1, 2 |
Epithelium, stroma |
Angiogenic factor |
Disease prognosis neg correlation |
van Moorselaar RJ, 2002) |
|
bFGF (FGF-2) |
Stroma |
FGF-2R |
Epithelium, stroma |
Angiogenic factor |
Disease prognosis pos correlation |
van Moorselaar RJ, 2002) |
|
HGF/SF |
Stroma |
c-met |
Epithelium |
Stimulates cell growth |
Disease progression pos correlation |
Knudsen BS, 2004; Lail-Trecker M, 1998 |
|
TGF-beta |
Epithelium |
TGF-beta I-III receptors |
Stroma |
Induces apoptosis, increases angiogenesis stimulates stromal but inhibits epithelial cell growth |
Augmented expression at androgen withdrawal |
Bachman, 2004 and 2005; Yingling, 2004 |
|
IGF-I |
Stroma |
IGF-IR |
Epithelium, stroma |
Stimulates cell growth, blocks apoptosis |
Up-regulation at disease Progression |
Djavan B, 2001; Rubin, 2003 |
|
IL-6 |
Epithelium, stroma |
IL-6R, sIL-6R |
Epithelium, stroma |
Promotes differentiation and apoptosis inhibition |
Increasing IL-6 signaling during disease progression |
Edwards, 2005a and b; Hideshima., 2005; Royuela M, 2004 |
|
(KGF) (FGF-7) |
Stroma |
Gp130 KGF-R |
Epithelium |
Stimulates cell Growth |
Stromal KGF expression responded to androgen |
Planz, 1999 and 2004 |
Recruitment of neovascular endothelial cells to proliferating cancer cells is thought to be required for the maintenance and stimulation of tumour growth, and is mediated by vascular endothelial growth factor (VEGF) and its receptors. VEGF has been shown to be secreted by both glandular and surrounding stromal cells, and VEGF expression can be modulated by a number of treatments including androgen ablation, finasteride, and thalidomide .
High levels of IL-6 secretion from prostate fibroblasts, smooth muscle cells, and tumour cells themselves, are thought to be a mechanism of ligand independent activation of AR activation via the PI3K-Akt, STAT3 and MAPK pathways in PCa . Interference with IL-6 signaling is a potential means of modulating the growth of advanced prostate cancer. Importantly, IL-6 is secreted by bone marrow stromal cells (BMSCs), and this secretion is further augmented by direct interaction between tumour cells and BMSCs . Studies using an anti-IL-6 monoclonal antibody have shown tumoricidal effects in a murine model
The insulin-like growth factor-I (IGF-I) pathway is involved with malignant transformation in various tissues. In prostate cancer, it has been proposed that IGF-1 induces ligand-independent activation of the androgen receptor and enhances the expression of matrix metalloproteinase-2 and urokinase plasminogen activator (see next section, insoluble cell signalling). Progression to androgen independence has also been linked to deregulation of the IGF-1-IGF-1-receptor axis (76). Manipulations of the IGF axis have shown therapeutic potential, for example, antisense RNA to IGF-I receptor inhibits prostate cancer proliferation and invasion, while increasing IGF binding protein 3 expression induces cell death .
HGS/SF, EGF and bFGF are all members of a large family of heparin bound growth factors. Hepatocyte growth factor/scatter factor (HGF/SF) and its receptor, the c-met proto-oncogene, were shown to be predominantly expressed by localized and metastatic prostate cancer. Experimental evidence suggests that HGF/SF and c-met downstream signaling may regulate prostate cancer growth and metastasis through enhanced IL-6, androgen receptor, extracellular matrix and integrin interaction . Epidermal growth factor (EGF) its receptor (EGFR), and other family members, erbB2/neu, erbB3 and erbB4, are known to have a role in prostate cancer progression through their interactions with a broad spectrum of soluble factors and their downstream converging signaling pathways (81, 82).
Transforming growth factor beta type 1 is a ubiquitous cytokine originally named for its ability to transform fibrolasts in culture (83). TGF-beta both inhibits the growth of normal epithelial cells, but paradoxically can induce cancer cell proliferation and promote an Epithelial to Mesenchymal transition (EMT). The mammalian TGF-beta family has 3 subtypes, and can bind to three cell surface receptors (type I, II and III). The type II receptor functions as a tumor repressor gene, and absence of this receptor results in resistance to the growth inhibitory effects of TGF-beta. This growth factor also stimulates angiogenesis, extracellular matrix turnover and host immune surveillance, although the receptors mediating this function have not been well defined (84). TGF-beta specific inhibitors have been developed and have been proposed to have utility in treatment of progressive prostate cancer .
At the stromal /tumor cell interface of adenocarcinomas, there is a noticeable derangement of Extracellular Matrix (ECM), due mainly to the activity of enzymes derived from the host stroma. (72). However, the mechanisms of this breakdown of ECM are unknown, and whether cancer cell invasion develops before or after interacting with host stroma, or whether stroma response is subject to reciprocal regulation by cancer cells remains undetermined. De Wever, and Mareel (2003)(86) suggest that the maintenance of epithelial homeostasis requires the participation of stroma, and therefore it is likely that stromal changes are subsequent to epithelial aberrations. ECM and its degradative products could signal to cancer cells through their cell surface integrin or non-integrin associated receptors. Subsequently, cell behaviours, such as cell polarity, secretion, adhesion, motility and invasion, and integrated cell functions, such as proliferation, differentiation and survival could be affected.
Integrins are a family of heterodimeric, transmembrane receptors that mediate the attachment of cells to the surrounding ECM and function as sensors of the environment. Alterations in integrin expression and signaling have been implicated in many aspects of tumorigenesis and metastasis including cell survival, migration, and invasion. In prostate cancer, the progression from normal to metastatic cells is accompanied by changes in the repertoire of integrins expressed and up-regulation of key adhesion-dependent signaling pathways . Important mediators of Integrin signaling include the adhesion kinases FAK (focal adhesion kinase) PAK (p21 activated kinase) and MAPK (a serine/threonine kinase) . Potentially, Integrins may have more complex roles by coordinating their actions with metalloproteinases and serine proteases, which together may increase cancer cell invasion and migration into secondary sites of cancer growth (72).
Overexpression and increased phosphorylation of FAK has been shown in a variety of tumours and these changes correspond to a malignant and metastatic phenotype . FAK has been shown to promote cell surface expression of metalloproteinases, particularly MT1-MMP. This results in activation of other MMP’s – particularly MMP2 enabling ECM degradation and tumour cell invasion .
Matrix metalloproteinases (MMPs) are involved in tumour invasion and metastasis in various malignancies. MMP-2 and MMP-9 are capable of digesting collagen type IV, a significant component of basement membranes, and have been implicated in prostate cancer progression and metastases ). Preliminary studies show the stimulation of protease induced receptors produced increased levels of MMP-2 and MMP-9 activity in prostate cancer cell lines, indicating their potential role in the metastasis of prostate cancer cells .
A crucial role for MMP-9 has been demonstrated in the colonization of bone by prostate cancer cells . Net MMP-9 activity in bone tissues peaked 2 weeks after injection of prostate cancer cells (PC-3), coinciding with a wave of osteoclast recruitment. In vitro, co-culture of PC3 cells with bone tissue led to activation of pro-MMP-9 and increased in secreted MMP-9 activity. Activation of pro-MMP-9 was prevented by metalloprotease inhibitors but not by inhibitors of other classes of proteases. The authors concluded that their data suggested that osteoclast-derived MMP-9 may represent a potential therapeutic target in bone metastasis .
MMP-2 has been implicated in initiation of metastases, that is in promoting the movement of tumour cells from the primary lesion into the lymphatic or circulatory systems , but does not appear to have a role in bone metasteses colonisation . Recently inhibition of MMP-2 by genistin, a form of dietary soy, via MAPK and TGF-beta, was demonstrated. This study suggested a physiological role for genestin and confirmed epidemiological studies which demonstrated that dietary intake of genisten was associated with lower rates of metastatic prostate cancer .
The plasminogen plasmin proteolytic cascade is a multi-functional pathway which facilitates a spectrum of biological processes including ECM remodelling during wound healing and tumour invasion, and metastasis. The urokinase-type plasminogen activator (uPA) and its receptor (uPAR), initiate this cascade by converting plasminogen to plasmin. Plasmin subsequently degrades a range of ECM components and activates MMPs (see above) (95-98). Over expression of uPA or uPAR is a feature of a number of malignancies, including prostate, and is correlated with tumour progression and metastasis. In contrast, inhibition of expression of uPA or uPAR or inhibition of uPA and uPAR interaction leads to a reduction in the invasive and metastatic capacity of many tumors (96, 98-100).
Kallikreins (KLKs) are highly conserved serine proteases that play key roles in a variety of physiological and pathological processes . Possibly the best known KLK with respect to prostate cancer is KLK3 or PSA. Recently, two other androgen regulated KLK’s have been shown to have altered expression in PCas, KLK2 and KLK4 . The protein product of KLK2, (glandular kallikrein; hGK2) is secreted in ejaculate, and like PSA has been shown to have a number of substrates (105). In particular, KLK2 can proteolyse several IGFB’s (IGFB-2, -3, -4 and -5) more efficiently and at lower concentrations than PSA, and potentially may stimulate the IGF axis. KLK4 may also play a role in the IGF axis by activating pro-PSA/KLK3 which induces IGF activity by cleaving IGFB-3. PSA (KLK3) also cleaves latent pro-transforming growth factor-beta (TGF-beta) to active TGF, thereby regulating prostatic cell growth and bone homeostatis.
In addition, it has also been suggested that PSA also plays a role in degradation of laminin (the ECM glycoprotein), stimulating cell invasion through the ECM. Both KLK2 and KLK4 can activate pro-uPA to active uPA, initiating the plasminogen plasmin proteolytic cascade (see above). KLK2, can also inactivate plasminogen activator inhibitor-1, repressing regulatory control of the plasminogen plasmin proteolytic cascade. (105)
Overexpression studies have shown that KLK4 induces transcriptional repression of E-cadherin, with associated increase in vimentin . Veveris-Lowe et al (2005) (106) concluded that the loss of E-cadherin and associated increase in vimentin are indicative of EMT and that KLK4, may have a functional role in the progression of prostate cancer through promotion of tumour cell migration.
One of the challenges in diagnosing prostate cancer is the heterogeneity of the disease phenotypically. As described above, alterations occur in cancer cells at many levels, both pre and post transcriptionally (in the genome and transcribed RNA), and post translational protein modifications (such as phosphorylations).
The quest for reliable prostate cancer markers to identify cancer cells in blood, bone marrow, urine, prostatic tissue itself and semen, has embraced a number of candidate biomarkers in addition to PSA. Many methods have been used to identify potential markers for prostate cancer, including analysis of gene expression (microarrays and quantitative PCR) and protein marker (histopathology, Mass-spectrophotometry) changes between disease and non-disease states.
Interestingly it has been noted that the genes, their transcription rates and subsequent proteins altered in prostate cancer can be grouped into a number of cellular pathways. These include cell adhesion, cell-cycle regulation, cell signaling, angiogenesis and apoptosis. Some of these proteins and genes have been discussed above, and therefore only their applicability as diagnostic/prognostic or therapeutic targets will be discussed in this section.
Table 2. Summary of molecular aberrations in prostate Cancer (Modified from Quinn et al., European Journal of Cancer 41 (2005) 858–887 [107])
|
Process |
Key molecules/markers |
References |
|---|---|---|
|
Apoptosis |
p53, Bcl-2, Clusterin |
Downing SR, 2003 Augustin, 2003; Rubio, 2005; Scaltriti, 2004a; Scaltriti, 2004b |
|
Signal transduction |
TGF-beta, KGF, EGF, Caveolins AR |
Chung, 2005; Williams, 2005; Yang; 2004 Edwards, 2005a and b |
|
Cell cycle regulation |
c-Myc, p16INK4A, p27KIP1, pRb, apoptotic index, Ki67 p53, EZH2 |
Quinn, 2005; Epstein 2005; Verambally, 2002 |
|
Cell adhesion and cohesion |
E-cadherin, alpha-catenin, delta-catenin, metalloproteinases, kallikreins, CD151 |
Quinn et al, 2005; Burger, 2002; Haese A, 2005; Kurek R, 2004; Lintula S, 2005; Steuber T, 2005; Stephenson et al, 1999; Ang, 2004 |
|
Angiogenesis |
VEGF, VEGF receptors, nitric oxide, PSMA |
Quinn et al, 2005; Burger, 2002; Chang, 2004 |
|
Other Molecular Markers |
AMACR; DD3/PCA3; pCGEM-1; Hepsin; PSCA |
Bussemakers, 1999; Landers, 2005; ; Schalken et al., 2005; Kumar-Sinha, 2004; Jiang, 2004; Srikantan V, 2002; ; Zhigang, 2005; |
As previously discussed, cell-cell signaling is mediated by a number of soluble and insoluble proteins and factors. These include growth factors, such as TGF-beta, EGF and IGF, and cytokines such as Il-6. Several of these factors have utility as biomarkers, such as TGF-beta and, as mentioned above, inhibitors to several of these compounds have been developed (72).
Caveolins are major structural proteins of Caveolae, specialized plasma membrane invaginations that are abundant in smooth muscle cells, adipocytes, and endothelium, and act as regulators of signal transduction (108). In a number of tumour models (including prostate) caveolin-1 has been implicated in oncogenic cell transformation and subsequent metastasis. Studies on knock-out mice and in breast cancer models, indicate that caveolin-1 normally functions as a negative regulator of cell transformation and tumorigenesis. However, in prostate cancer caveolin-1 may function as a tumour promoter, potentially via both genetic and post-translational modifications (108). In addition, a recent study showed that c-Myc and caveolin-1 immunopositivity correlated positively with Gleason score (P = 0.0253) and positive surgical margin (P = 0.0006). Yang (109) found that the combination of positive c-Myc and caveolin-1 in patients with clinically confined prostate carcinoma was a significant prognostic marker for disease progression after surgery.
As described above, changes in the androgen receptor have been long associated with prostate cancer development. While alterations in the function of this gene appear to be critical for cancer progression, these alterations occur at many levels, both in the genome (such as point mutations) and also post transcriptional and translational (ie at the RNA and protein levels (62, 63). Therefore using detection of any one of these changes as a screening tool for prostate cancer has not yet been shown to be practicable.
Aberrant expression of cell-cell adhesion molecules (C-CAMs) is often associated with the development of tumours. Decreased expression of many C-CAMs including E-cadherin, have been associated with the progression of prostate cancer and several other types of neoplasm (107). While CD44, another cell adhesion molecule, has been associated with prostate cancer, there have been inconsistencies across the few studies as to the prognostic value of this marker . While the loss of these molecules may be useful in immuno-histopathologically- based diagnoses, molecular based profiling may depend more on overexpressed molecules. Therefore, the activity of MMP’s 2 and 9, uPA and uPAR, and KLK2, 3 (PSA) and 4 (see above) in interfering with cell adhesion, may prove to be more relevant
Both KLK2 mRNA and its protein product, hK2, have been examined for efficacy as stand alone biomarkers or in combinations with other known biomarkers such as PSA . Thus far none of the published studies has found KLK2 to be useful prognostically and it appears to have limited diagnostic value; only one study has found that KLK2 (measured by quantitative PCR) in combination with PSA distinguishes cancers from BPH . A recent study from China found that a functional C748T polymorphism in KLK2 may be associated with increased risk for developing prostate cancer. The frequency of the CC, CT and TT genotypes was 65.7%, 32.7% and 1.6% in patients with prostate cancer and 56.0%, 37.5% and 6.5%, respectively, in controls (p = 0.010). Therefore, C allele carriers (CC and CT genotypes) were at significantly higher risk for prostate cancer than TT homozygous subjects (p = 0.002) .
As noted previously, KLK4, has also been associated with prostate cancer (114). The KLK4 protein product, hK4, is the first member of the KLK family that is intracellularly localized , and KLK4 expression is regulated by androgens, oestrogen and progesterone in prostate cancer cells. In situ hybridization on normal and hyperplastic prostate samples indicated that KLK4 is predominantly expressed in the basal cells of the normal prostate gland and overexpressed in prostate cancer .
Some cell adhesion associated molecules such as CD151, a member of the tetraspanin family which interacts with integrins (see above), may prove to be relevant prognostically. CD151 plays a role as a link between extracellular matrix and intracellular structures, and increased protein levels of CD151 in well and moderately differentiated prostate cancers correlate with disease relapse subsequent to radical prostatectomy (115)
Apoptosis or programmed cell death is part of cell growth and cycling in normal and benign cells. In cancers, key regulators of apoptosis can be pro-apoptotic or anti-apoptotic such as p53 and Bcl-2 respectively, show abnormal function and expression. While there are many genes and proteins which fall into this category, many are also cell cycle regulators (such as p21 and p16) and will be discussed below.
The pro-apoptotic protein p53 regulates transcription of genes required for G1-phase growth arrest of cells in response to DNA damage. Mutant p53 protein accumulation in malignant cell nuclei has been shown to be a poor prognostic indicator in several human carcinomas including breast, lung and colorectal (107). Mutations in p53 have been shown to be a common event in early stage, organ-confined prostate cancer and the loss of p53 function via expression of viral or cellular oncoproteins also seems common (116). A number of studies have reported that p53 nuclear accumulation in ≥ 20% of tumour cells is adversely prognostic (117). Although using p53 as a diagnostic marker has been debated due to the heterogeneity of expression within tumours, Quinn et al (2000 & 2005) (107, 117), propose that p53 has great potential as a prognostic marker as metastatic, recurrent and androgen resistant cancers show higher number of cells with p53 immunoreactivity compared with primary tumours.
Bcl-2 was initially identified as an apoptosis-inhibiting proto-oncogene in B-cell lymphomas. Its value as a diagnostic marker is unclear with some groups showing that only limited numbers of prostatic tumours express Bcl-2 (118, 119). However, other researchers have shown correlation with Bcl-2 and poor prognostic outcomes, with increased numbers of high-grade and metastatic tumours having Bcl-2 immunoreactivity . Also Bcl-2 overexpression in tumours has been associated with resistance to radiotherapy (123). Subsequently many scientists are looking at Bcl-2 inhibitors (such as transgenes, and antisense RNA oligonucleotides) as a means of sensitizing Bcl-2 expressing tumours to chemo and radiotherapies (125).
Another study showed that testosterone-repressed prostate message-2 (TRPM-2), also known as clusterin or sulfated glycoprotein-2, was elevated following androgen withdrawal in both normal and malignant tissues . In prostate adenocarcinoma, TRPM-2/clusterin expression may be useful as both a diagnostic and prognostic marker, with increased TRPM-2/clusterin protein expression evident in prostate cancer (96%) compared with BPH (73%) and normal prostate epithelium (17%) . Pins et al (2004) (129) indicated that TRPM-2/clusterin immunoreactivity in stromal cells surrounding the tumour epithelium predicted PSA relapse but staining within the primary tumour epithelium was not prognostic. The anti-apoptotic function of TRPM-2/clusterin is well documented, both by overexpression studies and through the activities of TRPM-2/clusterin specific inhibitors (124). However, some studies have shown pro-apoptotic functions of TRPM-2/clusterin in PC-3 androgen-independent prostate cancer cells. Cells overexpressing an intracellular, non secreted form of TRPM-2/clusterin showed signal-independent nuclear localization of the protein - leading to G2-M phase blockade followed by caspase-dependent apoptosis (127, 131). While TRPM-2/clusterin is an attractive target for theraputics, caution is warranted as it seems to have a number of functions in cell cycle and apoptosis, with the nuclear form (nClu) being proapoptotic while the secreted form (sClu) has prosurvival effects (131).
Genetic aberrations in the control of progression in the cell cycle are present in most human cancers (107). There are many molecules involved in this process whose expression is altered in prostate cancer, including the cyclin family, RBp (retinoblastoma protein), p16, p21, p27, p53, Smad4, FHIT, and PTEN/MMAC1. Increased expression of p16 and p21 has been associated with poorer prognostic outcomes, and while loss of RBp and p27 may also have some prognostic value, further studies are warranted (107, 110).
Strong expression of EZH2, a catalytic subunit of the polycomb repressor complex 2, in clinically localized prostate cancer is related to poor prognosis . EZH2 is also overexpressed in hormone refractory prostate cancers (133). Regulation of EZH2 is controlled by the E2F3 transcription factor and recently Foster (2004) (134) showed that nuclear expression of E2F3 in 20% or more of prostate epithelial cells is also an indicator of an unfavourable clinical outcome.
As discussed above, angiogenesis or blood vessel growth is an essential factor in cancer growth and progression (135). A key component in angiogenesis, VEGF, is highly expressed in most prostate cancers and has value prognostically (107). A number of VEGF and VEGF-R antibodies and peptide antagonists have been developed with the specific goal of targeting the neovasculature and growing cancer cells. This targeting pathway may become highly important since hypoxia is a known important factor that induces VEGF production .
Another molecule whose expression has been associated with angiogenesis is PSMA (prostate-specific membrane antigen) which is expressed by both tumor epithelium and tumour associated endothelial cells and neovasculature (137). A number of researchers, (138-140)have made a series of antibodies against the external domain of this protein, which have been used both pre-clinically and clinically for the diagnosis and therapy of prostate cancer. Chromosomal localization of PSMA gene, however, is controversial. It has been mapped to two regions, chromosome 11p11-12 and 11q14 and it has been proposed that a PSMA-like gene may exist in one of the two chromosomal regions through the process of gene duplication. Our research group has shown that a specific PSMA transcript has applicability as a biomarker for prostate cancer, particularly when used in combination with other gene transcripts (such as DD3/PCA3) (143, 144). PSMA may function as a ligand internalising receptor, an enzyme playing a role in nutrient uptake, and a peptidase involved in signal transduction in prostate epithelial cells (145).
A number of other gene transcripts have been identified which may have utility as diagnostic and prognostic markers that do not fit into the catergories discussed above. These include Hepsin a transmembrane serine protease, AMACR, (alpha-methylacyl-CoA racemase), PSCA (prostate Stem cell antigen) and the non-coding RNA’s DD3 and PCGEM-1.
Hepsin, a type II transmembrane serine protease is differentially expressed in prostate cancer compared with normal and BPH affected prostate tissue (144, 146). Interestingly, in-vitro studies in prostate cell lines found that Hepsin overexpression had growth inhibitory effects . More recently a model for Hepsin and tumour progression was found when the soluble form of Hepsin was found to activate HGF (hepatocyte Growth Factor) . Potentially, antagonists to Hepsin and subsequently HGF activation could be useful therapeutically.
Alpha-methylacyl-CoA racemase (AMACR) overexpression in prostate and other cancer tissues has been well characterized at both the mRNA and protein levels (151-153). This enzyme functions in the peroxisomal beta oxidation of branched-chain fatty acid molecules and has been implicated in the link between high meat high fat diets and the increased incidence of prostate cancer observed by many epidemiological studies (154). AMACR immunohistochemistry is being used in conjunction with normal Haemotoxylin and Eosin staining to aid in the diagnosis of prostate cancer histologically (128, 155, 156). As with many of the markers already mentioned clinicians are finding that using AMACR in conjunction with other prostatic markers gives better diagnostic results than AMACR alone (J153, 157). Molinie (2004)(157) showed that basal cells of normal prostatic glands stained with p63 in 100% of cases, while carcinomas had a p63-/AMACR+ profile, PIN were p63+/AMACR+, and benign lesions were p63+/AMACR-. Recently a number of splice variants of the AMACR transcript have been described which may be relevant for strategies targeting AMACR expression such as RNA antisense oligonucleotides (158).
PSCA was identified as a cell surface antigen expressed by prostate cancer cells and is regulated by the androgen receptor ). It is central to the development of the prostate gland and could provide a new diagnostic and therapeutic target for PCa (161). PSCA overexpression has more recently been shown in pancreatic and urothelial tumor models (162-164) and has been proposed as a target antigen with immuno-based therapeutics (165, 166).
DD3/PCA3 was identified by differential display technology in 1999 as a non-coding RNA highly specific to prostate cancer . Subsequently a number of researchers have confirmed the over-expression of DD3 (144,168) in a number of different cohorts. More recently the uPM3 (Bostwick) test has utilized this detection of this RNA in a PCR based assay in urines of prostate cancer patients. Two independent studies showed significant improvements in detecting cancer compared with the use of PSA alone (169, 170). Tinzl (2004)(169) reported 82% sensitivity, 76% specificity for the uPM3 assay compared to 98% sensitivity, 5% specificity, for tPSA (at a cutoff of 2.5 ng/ml). In the tPSA categories <4, 4-10 and >10 ng/ml sensitivity was 73%, 84% and 84% and specificity was 61%, 80% and 70%, respectively (169). The Canadian-based study (170) sampled 517 patients undergoing biopsy at five centres, for which 86% had an assessable sample. The overall uPM3 sensitivity and specificity in this sample group was 66% and 89%, respectively. Once again, in the tPSA categories <4, 4-10 and >10 ng/ml, the sensitivity of the uPM3 assay was 74%, 58% and 79% with specificity of 91%, 91% and 80% respectively. The positive predictive value of uPM3 was 75% compared with 38% for total PSA. (170).
PCGEM1, also a putative non-coding RNA was also identified by differential display analysis of paired normal and prostate cancer tissues, with subsequent Northern blot analysis of tissues showing that PCGEM1 was expressed exclusively in the human prostate. In-situ analysis showed tumor associated overexpression in 84% of prostate cancer patients while reverse transcription PCR assays revealed tumour-associated overexpression in 56% of patients . Interestingly PCGEM1 over-expression has been shown to be significantly higher in prostate cancer cells of African-American men than in Caucasian-American men (P=0.0002). In addition, ‘normal’ prostate epithelial cells from prostate cancer patients with a family history of prostate cancer also displayed increased PCGEM1 expression (P=0.0400). A physiological role for PCGEM1 in cell growth regulation has been suggested, with cells transfected with PCGEM1 displaying cell proliferation and an increase in colony formation .
Recently the focus of some studies has changed from identification of individual markers to utilising combinations of known prostate cancer-specific markers as predictors of disease recurrence after treatment with curative intent (172, 173). In our laboratory we have investigated the utility of using combinations of biomarkers in a PCR based diagnostic assay for prostate cancer. We found that using a combination of Hepsin, DD3 and PSMA allowed us to distinguish 100% of prostate tumour from BPH tissues (144). As with diagnosis, it has been proposed that a number of genes can be used to build a “fingerprint” of an aggressive tumour . The challenge remaining is how to apply this information in biologically relevant samples such as serum, ejaculate and urine sediments.
With the development of methods that allow large scale gene expression profiling, such as microarrays and quantitative RT-PCR, the list of genetic alterations in prostate cancer cells has increased dramatically. Now the same technology is being applied to proteins (proteomics). These advances in technology should allow the scientist to determine that genetic changes translate into the proteome and identify post-translational modifications which are biologically significant in cancers. Although this chapter has attempted to summarise the current literature, this area of science is ever-expanding and subsequently there are many other biomarkers for prostate canceer, and other cancers which may be relevant in PCa, that were not discussed. Over the past decade there has been an explosion of research into the basic science of prostate cancer. Consequently, we are developing new paradigms for understanding the natural history of the disease as well as creating novel approaches to therapy. Later in this chapter, some of these advances will be discussed in further detail.
Currently, surgery and radiation therapy are the conventionally accepted primary treatment modalities for localized prostate cancer. However, with the advent of new molecular-based therapies, combination therapy using surgery and/or radiation with these novel agents has been proposed as a method of improving therapeutic outcomes. More importantly, molecular-based therapies are being studied in phase I, II, and II trials in men with advanced and hormone-refractory prostate cancer, for which no curative therapies currently exist.