Neuroendocrine Differentiation in Prostate Cancer
Review of recent literature with an emphasis on recent developments and basic mechanisms
The epithelial components of the benign prostate consist of luminal secretory cells and basal cells that can be identified on H&E-stained sections under light microscope. In addition, there is a third minor component of neuroendocrine (NE) cells which can only be identified by electron microscopy or immunohistochemical staining with antibodies for NE markers (chromogranin, synaptophysin, neuron-specific enolase) or NE products. Prostatic NE cells are intraglandular and intraductal cells with hybrid epithelial, neural and endocrine features. They express and secrete serotonin and numerous peptides/ neuropeptides. They are generally widely scattered throughout the prostate with only an occasional cell per gland/ duct, but are most consistently found in the periurethral ducts and verumontanum. Among the different zones of human prostate, transition zone and peripheral zone have more abundant NE cells than central zone, suggesting their potential involvement in benign prostatic hyperplasia and prostate cancer, respectively (Santamaria et al, 2002). A small percentage of human prostates contain numerous NE cells.
The prostate NE cells are of the open and closed types. The cells of the open type have an apical cytoplasmic process which extends to the lumen and has long specialized surface microvilli; while those of the closed type are surrounded by other epithelial cells and do not have direct contact with the lumen. Both types of NE cells have long branching dendrite-like processes which extend between nearby epithelial cells. Ultrastructural studies have shown a wide range of neurosecretory granule morphologies which correlate with the large number of known secretory products, including serotonin, histamine, chromogranin A and other members of the chromogranin family of peptides, calcitonin, calcitonin gene-related peptide, katacalcin, neuropeptide Y, vasoactive intestinal peptide (VIP), bombesin/gastrin releasing peptide (GRP), somatostatin, alpha-human chorionic gonadotropin (alphaHCG), parathyroid hormone-related protein (PTHrP), thyroid stimulating hormone-like peptide, cholecystokinin, adrenomedullin and vascular endothelial growth factor (VEGF) (reviewed in Huang and di Sant’Agnese, 2002). Some of these NE cell products have been detected in seminal fluid raising the possibility that they may be actively secreted into the seminal fluid and regulate sperm function or female genital tract activity. Receptors for some of the NE products have been localized to the prostate and/or prostate cancer including serotonin (5HT1a) (Abdul et al, 1994; Dizeyi et. al. 2004), bombesin/GRP (GRPR) (Aprikian et al, 1996; Sun et al, 2000; Markwalder and Reubi, 1999; Reubi et al, 2002), neurotensin (Seethalakshmi et al, 1997), somatostatin (SSTR1-5) (Halmos et. al. 2000; Berruti et al, 2001; Dizeyi et. al. 2002; Hansson et. al. 2002), cholecystokinin (Petit et al, 2001), Neuropeptide Y (Magni and Motta, 2001) and calcitonin (Wu et al, 1996). Hence, it is proposed that the NE cells may regulate the growth, differentiation and secretory activity of the prostatic epithelium, possibly through a paracrine mechanism. On the other hand, the neural network may play a role in regulating the activity of the NE cells, the contents of the glandular lumen or endocrine, paracrine or autocrine signaling.
Neuroendocrine Differentiation in Prostate Cancer
NE differentiation (NED) is also present in prostate cancer. Some carcinomas are completely differentiated along NE lines, most notably carcinoid tumors and small cell carcinomas which are very rare. Somewhat more common are mixed tumors of conventional adenocarcinoma with a component of small cell carcinoma or carcinoid tumor. A recent report has proposed a category of large cell neuroendocrine carcinoma of the prostate (Evans et al, 2004). Generally, the term NED in prostate cancer refers to the presence of scattered NE cells singly or in small nests in conventional prostatic adenocarcinomas.
Types of Neuroendocrine Differentiation
Small Cell Carcinoma
Small cell carcinoma of the prostate is rather rare, particularly in its pure form, and accounts for no more than 1 percent of all carcinomas of the prostate. They are aggressive tumors which often present at advanced stages or as metastatic diseases (Erasmus et al, 2002) and occasionally associated with paraneoplastic syndromes (Kawai et al, 2003). Some small cell carcinomas represent recurrent tumors after hormonal therapy for conventional adenocarcinomas of the prostate (Tanaka et al, 2001; Miyoshi et al, 2001; Spieth et al, 2002). More commonly, small cell carcinoma is present as a component of mixed tumors which also contain a component of conventional adenocarcinoma. Histologically, small cell carcinomas of the prostate are similar to the more common small cell carcinomas of the lung. They are characterized by the following features: 1) solid, sheet-like growth pattern, often with areas of tumor necrosis; 2) small, round to spindle cells with scant cytoplasm, high nuclear/cytoplasmic ration and ill-defined borders; 3) hyperchromatic nuclei with finely granular chromatin and nuclear molding; 4) absent or inconspicuous nucleoli, and 5) high mitotic rate (> 10/10 HPF's). (Figure 3A).
The solid growth pattern of small cell carcinoma of the prostate makes distinction of such tumors from Gleason grade 5 adenocarcinomas difficult at times, for which immunohistochemical study may be of help. Small cell carcinomas are often positive for NE markers chromogranin-A (Figure 3B), synaptophysin and NSE although one or more of these markers may be negative in any given case. Like small cell carcinomas of the lung, tumor cells often show dot-like cytokeratin staining pattern and are often positive for TTF-1. In contrast to prostatic adenocarcinoma, tumor cells of small cell carcinoma are usually negative for androgen receptor and PSA but exceptions exist (Kawai et al, 2003). It is important to keep in mind that immunohistochemical profile varies from case to case and morphology remains the gold standard for pathological diagnosis.
The prognosis for small cell carcinomas is poor and the disease is usually rapidly fatal (Papandreou et al, 2002). Hormonal therapy is not effective in treating such tumors and neither is surgery. The general response to chemotherapy, the main form of therapy for small cell carcinoma, is some initial response but progressing to a rather rapid downhill course (Moore et al, 1992, Rubinstein et al, 1997, Helpap, 2002).
Carcinoid tumors have also been reported in the prostate, which are exceedingly rare. These tumors show complete NE differentiation and have morphologic features similar to carcinoid tumors of the lung or GI tract. In comparison to small cell carcinomas, the tumor cells have more abundant cytoplasm, rare mitotic figures and no tumor necrosis. Like small cell carcinomas, carcinoid tumors often exist as a component of mixed tumors also containing conventional adenocarcinoma (Ghannoum et al, 2004).
Focal NED in Prostate Cancer
NED in prostate cancer generally refers to the presence of NE cells focally in otherwise typical conventional adenocarcinomas (Figure 4). The NE cells are indistinguishable from the non-NE cancer cells morphologically and are usually identified by immunohistochemical study for NE markers or less commonly, by electron microscopy. Chromogranin-A is the most commonly used marker and is considered to be sensitive and specific. All carcinomas of the prostate have at least some NED (Abrahamsson et al, 1987) when multiple generic NE markers and/or specific peptides/neuropeptides are used and tissue preparation is controlled. About 5-10% of prostatic carcinomas have rather extensive multifocal NED. The significance of NED as an independent prognostic factor in androgen responsive prostate cancer is controversial. Some studies showed independent prognostic significance (Dema et al, 1996; Weinstein et al, 1996; Theodorescu et al, 1997; Cohen et al, 1991; Bollito et al, 2001; Yu et al, 2001) while many others did not (Bohrer and Schmoll, 1993; Aprikian et al, 1994: Cohen et al, 1994; Allen et al, 1995; Noordzij et al, 1995; Bubendorf et al, 1996; McWilliam et al, 1997; Speights et al, 1997; Abrahamsson et al, 1998; Casella et al, 1998; Segawa et al, 2001; Bostwick et. al. 2002; Steineck et al, 2002). There is more consistent evidence based on immunohistochemical and serologic studies that NE differentiation is a prognostic factor in androgen independent prostate cancer (Krijnen et al, 1997; Jiborn et al, 1998; Ischia et al, 2000; Berruti et al, 2000). NED increases in high grade/high stage tumors (Abrahamsson et al, 1989; Bohrer and Schmoll, 1993) and, particularly, in androgen deprived (Ahlgren, 2000; Ismail et al, 2002) and androgen independent tumors (Jiborn et al, 1998; Ito et al, 2001; Hirano et al, 2004), although divergent findings have been reported (Kollermann and Helpap, 2001; Li et al, 2003). Consistent with these findings, serum chromogranin-A levels are increased in patients with advanced, androgen independent cancers (Deftos et al, 1996; Cussenot et al, 1996; Berruti et al, 2000; Ferrero-Pous et al, 2001; Isshiki et al, 2002). Measuring serum chromogranin A levels may be of value in prostate cancer patients with false negative PSA or %F-PSA (Ahel et al, 2001) and can be used to monitor treatment response (Zaky Ahel et al, 2001). In patients with advanced prostate cancer receiving hormonal therapy, increase in the serum levels of chromogranin-A may precede PSA elevation and signal treatment failure (Chuang et al, 2003) and correlates with bone metastasis (Tarle et al, 2002). Intermittent administration of complete androgen deprivation therapy significantly reduces the increase in serum chromogranin-A levels in comparison to continuous therapy (Sciarra et al, 2003). The serum levels of NSE may also have prognostic significance (Kamiya et al, 2003; Hvamstad et al, 2003). Other serum markers such as chromogranin B, secretoneurin, which is a proteolytic product of secretogranin II (chromogranin C), and gastrin-releasing peptide/ProGRP may serve as additional prognostic and/or diagnostic markers (Angelsen et al, 1997B; Ischia et al, 2000; Lilleby et al, 2001; Yashi et al, 2002, 2003; Nagakawa et al, 2002). Serum calcitonin appears to be a more specific marker for small cell carcinoma of the prostate (Sim et al, 1996).).
Androgen Deprivation and Neuroendocrine Differentiation
In addition to histologic studies, there is biochemical evidence that androgen deprivation may induce NE activity in prostate cancer. Neutral endopeptidase 24.11 (NEP) is a cell surface enzyme expressed by prostatic epithelial cells and functions to cleave and inactivate a variety of neuropeptides. Down-regulation of NEP after androgen deprivation may lead to increased secretion of neuropeptides such as neurotensin by the NE cells. Interestingly, only androgen-deprived tumor cells respond to the growth-promoting effect of neurotensin (Sehgal et al, 1994). The expression and catalytic activity of NEP are lost in androgen-independent but not androgen-dependent prostate cancer cell lines. In vivo, metastatic cancer cells from patients with androgen-independent prostate cancer commonly show decreased levels of NEP compared with those from patients with androgen-dependent prostate cancer. Growth of androgen-independent cancer cells is inhibited by overexpression of NEP or incubation with recombinant NEP (Papandreou et al, 1998).
Models of Neuroendocrine Differentiation
Animal models of prostate cancer recapitulate human diseases and provide more evidence suggesting the importance of NED in prostate cancer. In xenograft models of human prostate cancer, NED increases markedly after castration (Burchardt et al, 1999; Jongsma et al, 1999, 2000, 2002; de Pinieux et al, 2001) and precedes the emergence of increased cancer cell proliferation and progression to hormone-independent cancer (Huss et al, 2004). Extensive NED is also seen in an allograft model of androgen-independent prostate cancer (Masumori et al, 2004). In a transgenic mouse model of prostate cancer (TRAMP), the degree of NED correlates with the degree of tumor differentiation with the poorly differentiated tumors showing significantly more NED. Castration of TRAMP mice leads to aggressive and highly metastatic cancers in the majority of the cases, reflecting androgen-independent growth. NED was detected in the majority of the primary and metastatic tumors in such animals (Kaplan-Lefko et al, 2003). Animal models of NE/small cell prostate carcinoma have also been established and should prove useful in studying the molecular basis of NE phenotype (Garabedian et al, 1998; Hu et al, 2002; True et al, 2002).
Prostate Cancer Cell Lines
In vitro, LNCaP cells, an androgen-dependent cell line, can be induced to show NED by androgen deprivation or agents that increase intracellular levels of cAMP. The NE phenotype is rapidly lost upon withdrawal of inducing agents (Cox et al, 1999). These findings support the transdifferentiation model and suggest that the tumor NE cells may be derived from non-NE tumor cells. The caveat of this model is that NED of LNCaP cells is an “all or none” phenomenon and all cells acquire NE phenotype under appropriate culture conditions while NED in human cancers is focal, making interpretation of the in-vitro findings difficult. Alternatively, some investigators believe that NE cells may be derived from the same stem cell or pluripotent cell that give rise to luminal secretory cells (Bonkhoff and Remberger, 1996; Rumpold et al, 2002). In comparison to secretory epithelial cells, the NE cells, whether benign or malignant, generally do not express androgen receptor and PSA. They are considered terminally differentiated and post-mitotic (Bonkhoff et al, 1991, 1995, 1997). However, a recent report suggests that the malignant NE cells may possess proliferative activity (Huss et al, 2004). The origin of NE cells in prostate cancer is not entirely known. A population of transiently proliferating/amplifying intermediate cells has been identified which may be the common precursor for NE cells and other epithelial cells of the prostate (Xue et al, 1997; Schalken and van Leenders, 2003). Nonetheless, it is generally accepted that NE cells in prostate cancer are malignant since they are present in metastatic prostate cancers (Roudier et al, 2003, 2004).
Neuroendocrine Function in Prostate Carcinoma
The function of NED in prostate cancer has been extensively studied. Results from a recent study using xenograft mice and in-vitro assays suggest that NE cells may directly activate androgen-receptor, thus promoting growth of LNCaP tumor cells in the absence of androgen (Jin et al, 2004). Bombesin acts as a mitogen in prostate cancer and may do so through activation of the transcription factor Elk-1 and the immediate early gene c-fos (Xiao et al, 2002). In in-vitro assays, neuropeptides stimulate androgen-independent growth (Jongsma et al, 2000B) and the invasiveness of prostate cancer cell lines (Hoosein et al, 1993). Both protein tyrosine kinase and protein kinase C pathways may be required for the activity of neuropeptides (Aprikian et al, 1997; Allard et al, 2000; Salazar and Rozengurt, 1999; Sumitomo et al, 2000; Xiao et al, 2003). The activity of the type IV collagenases matrix metalloprotease (MMP) is up-regulated by neuropeptides (Sehgal and Thompson, 1998). MMPs are associated with a variety of biological activities, such as tumor invasion, metastasis, and angiogenesis. MT1-MMP protein and mRNA are expressed in androgen-independent PC-3, DU-145 and TSU-pr1 cells but not in the androgen-dependent LNCaP cells. GRP induces the expression of MT1-MMP protein in DU-145 cells and also increases Matrigel invasion by these cells (Nagakawa et al, 2000). High-grade tumors are more likely to express MMP-9 and bombesin than low-grade tumors (Ishimaru et al, 2002). Bombesin increases the expression of the proteolytic enzyme urokinase-type plasminogen activator (uPA) and plasminogen activator inhibitor-1 (PAI-1) and also stimulates secretion and activation of MMP-9 (Festuccia et al, 1998). Calcitonin affects growth and migration of certain prostate cancer cell lines and may play a role in the regulation of prostate cell growth and metastases, especially to the bone (Ritchie et al, 1997). Certain receptors for serotonin may be overexpressed in prostate cancer cells, particularly in high grade tumors, further supporting the hypothesis that NE products may promote androgen-independence of prostate cancer through a paracrine mechanism (Dizeyi et al, 2004).
NED may affect the apoptosis-resistance of prostate cancer.
The tissue levels of NSE correlate with Bcl-2, a major anti-apoptotic factor, and the Bcl-2-containing cancer cells are generally in close proximity to the NE cells (Segal et al, 1994), suggesting that NE cells may confer apoptosis-resistance to the neighboring cancer cells. Bombesin and calcitonin prevent apoptosis of prostate cancer cells in-vitro (Salido et al, 2000, 2004; Vilches et al, 2004). NE cells do not appear to undergo apoptosis (Bonkhoff et al, 1997; Fixemer et al, 2002) even though they are negative for Bcl-2 (Xue et al, 1997). They do express survivin, another anti-apoptotic factor (Xing et al, 2001), providing a molecular basis for the hypothesis that NE cells may endure stressful conditions and escape from apoptosis during cancer therapy.
NED may promote neovascularization of prostate cancer.
NE cells are the major producers of VEGF (Harper et al, 1996; Chevalier et al, 2002). VIP, which is produced by autonomic nerves and NE cancer cells of human prostate, stimulates NED and VEGF production in LNCaP cells (Collado et al, 2004). In benign prostatic tissue, there is negative to low level expression of VEGF. All prostate cancers stained positively for VEGF and staining intensity correlated with Gleason grade. Complete androgen block for three months before surgery decreased the level of VEGF and vascularization, except in the cell areas with NE features (Mazzucchelli et al, 2000). In radical prostatectomy specimens, there is a correlation between NE differentiation and neovascularization and both correlate with tumor grade and tumor stage. The number of NE cells was found to be the only predictor of neovascularization (Grobholz et al, 2000). Another study reported similar results and, in addition, found expression of VEGF to be significantly correlated with increased microvessel density, high tumor stage, poor differentiation and shortened disease-free survival (Borre et al, 2000). Bombesin stimulates expression of pro-angiogenic factors VEGF and IL-8 in PC-3 cells, possibly through NF-kappa B-dependent pathway (Levine et al, 2003). However, a recent study failed to show stimulation of proliferation, migration, or tube formation of human umbilical endothelial cells in-vitro by neurotensin and bombesin (Busby et al, 2004).
Molecular Mechanism of Neuroendocrine Differentiation
Androgen receptor may actively repress the NE phenotype (Wright et al, 2003), providing an explanation for the emergence of the NE phenotype when AR signaling is inhibited, such as in hormonally-treated cancers or in LNCaP cells cultured in androgen-deprived media. Abrogation of FGF signaling by expression of a truncated FGFR2iiib receptor in prostatic epithelium of transgenic mice promotes NED (Foster et al, 2002). The same is observed in expression of a constitutively activated form of the heterotrimeric G-protein subunit alpha (Regnauld et al, 2002). Interaction of IGF-binding protein-related protein 1 with a novel protein, neuroendocrine differentiation factor, results in NED of prostate cancer cells (Wilson et al, 2001). Inhibition of COX-2, a proinflammatory enzyme, induces NF-kappa B and NED of LNCaP C4-2b subline (Meyer-Siegler, 2001). NED of LNCaP cells can be induced by agents increasing intracytoplasmic levels of cAMP (Cox et al, 1999), IL-6 or IL-1 (Albrecht et al, 2004). PKA signaling pathway may be important for NED (Cox et al, 2000). NED of LNCaP cells induced by heparin-binding epidermal growth factor-like growth factor (HB-EGF) involves mitogen-activated protein kinase (MAPK) signaling pathway (Kim et al, 2002). Papaverine combined with prostaglandin E2 (PGE2) synergistically induces NE differentiation of LNCaP cells (Shimizu et al, 2000). NED of LNCaP cells is accompanied by overexpression of an alpha 1H (Cav3.2) T-type calcium channel (Mariot et al, 2002) and changes in intracytoplasmic calcium homeostasis (Vanoverberghe et al, 2004). Protein tyrosine phosphatase alpha, a receptor type protein tyrosine phosphatatse, (Zelivianski et al, 2001; Zhang et al, 2003) and mAsh1, a basic helix-loop-helix (bHLH) transcription factor (Hu et al, 2004) may be involved in NED of prostate cancer cells. Silibinin, a flavonoid antioxidant, induces G1 arrest and NED of LNCaP cells through increasing Rb level, decreasing Rb phosphorylation and inhibition of key cell cycle regulators (Zi and Agarwal, 1999; Tyagi et al, 2002).
Recently, it has been demonstrated that neuroendocrine tumor cells in human prostate cancer tissue express interleukin-8 (Huang et al, 2005). IL-8 was initially identified as a regulator of leukocyte recruitment and trafficking (Onuffer JJ et al, 1992) but has also been found to be a mitogenic (Inoue et al, 2000) and angiogenic factor (Koch et al, 1992). In patients with PC, serum IL-8 levels increase with progression of the disease (Veltri et al, 1999, Lehrer et al, 2004). The PC cell line, PC3, expresses and secretes IL-8 (Moore et al, 1999) and also expresses the IL-8 receptor CXCR1 and CXCR2 (Reiland, 1999). PC cell line LNCaP does not express IL-8, but selection of the cells in androgen-deprived media led to the emergence of a cell line which produces IL-8 and is more tumorigenic than the parental cells (Patel et al , 2000). A recent study showed that IL-8 promotes androgen-independent growth and migration of LNCaP cells (Lee et al, 2004), suggesting that IL-8 may facilitate transition of PC to an androgen-independent state.
IL-6-induced NE differentiation involves the protein tyrosine kinase pathway (Chung et al, 2000; Qiu et al, 1998), members of the JAK-STAT family (Spiotto and Chung, 2000; 2000B), induction of cyclin-dependent kinase (CDK) inhibitor p27 (Kip1) and inhibition of CDKs (Mori et al, 1999). Androgen and androgen receptor can regulate IL-6-mediated LNCaP cell NED via directly modulating the IL-6-PI3-kinase pathway (Xie et al, 2004). Interestingly, IL-6-induced NED of LNCaP cells may be qualitatively different from that induced by agents such as epinephrine and forskolin which cause rapid but reversible NE differentiation of LNCaP cells by increasing intracellular concentrations of cAMP. The process of IL-6-induced NE differentiation takes more time and is permanent (Wang et al, 2004). In addition, the behavior of IL-6-treated cells may depend on the concentration of IL-6 used. Long-term exposure of LNCaP cells to low concentrations of IL-6 (5 ng/ml) results in the emergence of a LNCaP variant with more aggressive growth properties in vitro and in vivo (Hobisch et al, 2001; Steiner et al, 2003) while culture of LNCaP cells in high concentrations of IL-6 (100 ng/ml) for 2 weeks leads to permanent NED and significant loss of the proliferative potential (Wang et al, 2004). Therefore, even though NED is considered to promote progression of prostate cancer to androgen-independent state, it is possible that complete differentiation of cancer cells along the NE pathway may actually inhibit proliferation, providing a novel strategy for developing anti-prostate cancer therapy.
Summary and Conclusions
NE cells constitute the third component of prostatic epithelial cells and may play a role in the normal development and function of benign prostatic tissue. They are also present in prostate cancers and their number and activity increase in high grade and androgen deprived-cancers, particularly in those that are androgen-independent. The origin of NE cells in prostate cancer is unclear and they may be derived from transformed stem cells or from non-NE cancer cells through transdifferentiation. In-vitro and in-vivo evidence suggest that the products of the NE cells may contribute to the emergence of androgen-independence by acting on the receptors present in non-NE cancer cells in a paracrine fashion. Identification of the key molecules required for NE differentiation or the crucial NE cell effectors that promote androgen-independent growth may provide novel targets that can be exploited to prevent the progression of prostate cancer to the hormone refractory state.