Several agents have been developed to target HDACs. These inhibitors exhibit pleiotropic effects including cell cycle arrest, DNA damage, autophagy, and apoptosis [9–11]. To date, there are at least 18 HDAC inhibitors (HDACi), divided into seven classes, being investigated in prostate cell lines or animal models [1]. In particular, vorinostat, entinostat, and panobinostat are being investigated in clinical trials.

Vorinostat, also named as suberoylanilide hydroxamic acid (SAHA), is a hydroxamate compound inhibiting class -I, -II, and -IV HDAC and has been approved to treat cutaneous T cell lymphoma by FDA in 2006 [1]. In prostate cancer therapy, it is able to induce cell death in vitro in LNCaP, LAPC4, CWR22, DU145 [12], and reduce CWR22 tumor burden in vivo [13]. Molecular analysis shows that vorinostat represses AR, PSA, and KLK2 transcription, but does not affect AR degradation [14]. Moreover, numerous studies displayed that vorinostat arrests cell cycle in G2/M, repressing EGFR expression, and has been reported to have synergistic effect with bicalutamide in DU145 [14].

Trichostatin A (TSA) is an antifungal antibiotic, specifically blocking mammalian class I and II HDACs [1]. One study showed that TSA induces cell death in LNCaP and CWR22R, but not in PC3 and DU145 [15]. Suenaga et al. suggested that TSA is capable of suppressing telomerase reverse transcriptase (hTERT) mRNA, which is responsible for maintaining telomere integrity, to restrain LNCaP and PC3 cells proliferation [16].

Panobinostat is a derivative of hydroxamic acid, the same class as SAHA. It has great activity against class -I, -II, and -IV HDAC at nanomolar ranges [1]. Some studies have indicated that panobinostat induces cell apoptosis in LNCaP [4], MYC-CaP/AS [17], MYC-CaP/CR [17], as well as PC3-AR expressing cells [18] and decreases tumor size of MYC-CaP [17], CWR22RV1 [4], PC3-AR [18] in vivo models. Several research teams have revealed that panobinostat blocks AR transcriptional activity, induces Caspase-3 activations, and attenuates ATM–Akt–ERK DNA damage pathway to cause cell death [4, 17, 18].

Romidepsin, also named as FK228, belongs to the class of bicyclic and capable of releasing a zinc-binding thiol in a cell to repress HDAC activity [1]. Sasakawa et al. have indicated that romidepsin inhibits PC3 and DU145 cell growth as well as tumor burden by inducing p21 and downregulating c-Myc expression [19]. Besides, the other group displayed that romidepsin can repress tumor volume in the 22Rv1 model and prolong survival rate [20].

Entinostat is a benzamide-based HDACi and targeting class I HDAC exclusively. It has been approved by FDA for ER-positive breast cancer treatment and investigated in LNCaP, PC3, as well as DU145 cells showing anti-tumor effect. [21]. Moreover, a study concluded that entinostat acts as a radiosensitizer to enhance radiation mediated DNA double strand breaks (DSB), as indicated by increased levels of gamma-H2AX [22]. Importantly, Qian et al. demonstrated that entinostat significantly prevents the progression of prostate tumorigenesis in TRAMP model [21].

The efficacy of HDAC inhibitors as single agents has been demonstrated in various prostate cancer cell lines and in vivo models. However, the development of CRPC results from multiple signaling pathways [24] (Table 20.2). For this reason, combining HDACi with other agents has been studied in preclinical PCa animal models and clinical trials in patients with PCa. Marrocco et al. combined vorinostat with the AR antagonist, bicalutamide, showing synergistic cell killing effect in the LNCaP cell line [14]. Also, Roklhin et al. displayed the synergistic effect of TSA and doxorubicin in LNCaP and CWR22R cells [15]. Moreover, due to the fact that panobinostat results in PI3K–AKT–mTOR pathway repression, Ellis et al. adopted Myc-CaP and PC3 models to demonstrate the greater therapeutic efficacy in vitro and in vivo by combining panobinostat with the specific mTORC1 inhibitor everolimus and the PI3K–mTOR dual inhibitor BEZ235 [17, 18].

The promising results from these in vitro and preclinical in vivo studies involving HDAC inhibitors in various PCa models are encouraging. These results have led to numerous clinical trials, investigating the potential of HDAC inhibitors as single agent or in combination with radiotherapy, bicalutamide, docetaxel, doxorubicin, and temsirolimus [25] (Table 20.3). An example of a promising clinical trial involved the treatment of patients with docetaxel and panobinostat demonstrating PSA reduction [26]. Although most trials are still ongoing and vorinostat single treatment in CRPC has been reported high toxicity, low dose of HDAC inhibitors or combining with other drugs may still provide an avenue to treat CRPC patients.

Like acetylation and deacetylation of histones, methylation of DNA and histones are reversible alterations that lead to stable inheritance of cellular phenotypes without any changes in the DNA sequence (Fig. 20.2). During DNA methylation process, the methyl group (–CH₃) is supplied by S-adenosylmethionine (SAM) and transferred to the 5′-carbon of a cytosine base of a CpG dinucleotide. Regions characterized by high frequency of CpG sites are defined as CpG islands and their methylation acts as a stable tag on gene promoters, leading to the formation of large-scale heterochromatic structures that silence the associated genes.

Histone methylation occurs on arginine and lysine residues of histone tails. The methylation is catalyzed by histone methyltransferase enzymes and a single lysine residue can be methylated up to three times leading to either gene repression and activation. Both DNA and histone methylation can drive to genetic instability and alteration in normal gene transcription and have been implicated in a variety of human cancers, including PCa [31]. For those reasons, these epigenetic modifications can potentially be used for the molecular classification, detection, risk assessment, and treatment in prostate cancer. In the following sections, the principal epigenetic molecules involved in PCa progression and their relative inhibitors will be presented.

Histones are the main structural components of chromosomes. These proteins assemble to form octamer and act as spools around which DNA is wrapped to form a nucleosome. Histones undergo a sophisticated pattern of posttranslational modifications, the histone code, that alter their interaction with DNA and nuclear proteins and are needed to regulate gene transcription. Histone methylation can affect chromatin remodeling and function depending on the specific amino acid being modified and the extent of methylation and can lead to both activation and repression (Fig. 20.3) [31]. The two histone methylation marks, therefore, associated with active transcription are the trimethylation of histone H3 at lysine 4 (H3K4Me3) and the trimethylation of histone H3 at lysine 36 (H3K36Me3). Among the repressive marks are the trimethylation of histone H3 at lysine 27, 9, and 20 (H3K27Me3, H3K9Me2/3, and H3K20Me3). Increasing evidence suggests that alterations in the histone code may play an important role during prostate tumorigenesis and the expression levels of histone methyltransferases are often altered in multiple tumor types. In 2009, two independent groups have shown that, in CRPC models, enhanced levels of the active histone marks H3K4me1/2/3 are selectively enriched at the enhancer regions of some AR-regulated cell cycle genes thus promoting CRPC cell growth and survival [43, 44]. Further, an increase in the repressive histone marks H3K27me3 has also been shown during prostate cancer progression, correlating with a poor clinical outcome. H3K27me3 increase is attributed to the overexpression of the Enhancer of Zeste Homolog 2 (EZH2) and leads to the silencing of a wide number of genes (such as tumor suppressor genes, GAS2, PIK3CG, and ADRB2), whose repression can induce cancer cell growth and invasion. EZH2 is a histone methyltransferase belonging to the polycomb repressive complex 2 (PRC2) and its overexpression has been seen in metastatic prostate cancer [45] following microRNA-101 deletion, a negative regulator [46]. Despite several studies have focused on PCR2-mediated repression as EZH2 oncogenic function, in 2012, Xu et al. have demonstrated that EZH2 overexpression in a model of androgen-independent PCa (LNCaP-abl) correlates with a switch in EZH2 activity from gene repressor to gene activator. EZH2 was found to be highly phosphorylated on a specific serine residue (S21) and, depending on its phosphorylation, able to interact and act as an AR transcriptional co-activator. The importance of phosphorylation at S21 has also been confirmed for the androgen-independent growth of the androgen-sensitive cell line, LNCaP [47]. Moreover, H3K27me3 levels are significantly decreased in CRPC, further supporting the idea of an EZH2 oncogenic activity independently of its Polycomb repressive function [47]. A deregulated expression of EZH2 has also been described in other solid tumors such as: bladder, gastric, lung, and hepatocellular carcinoma. Interestingly, recent work showed that EZH2 can function as an upstream activator of another HMTase, MMSET/NSD2. NSD2 mediates H3K36me2, a histone mark associated with active transcription, and its expression and function has been reported to be tightly related to EZH2 in most human cancers [48]. In CRPC, NSD2 is overexpressed and acts as a strong co-activator of NF-κB, promoting cell proliferation and survival and tumor growth [49].

Histone lysine demethylases (KDMs) are enzymes able to remove both repressive and activating histone marks (Table 20.5). KDMs are grouped in seven major classes, each one targeting a specific histone methylation site. Belonging to the first class, KDM1, are two isoforms of flavin-dependent demethylases (KDM1A and B, also known as LSD1 and LSD2) responsible for the demethylation of H3K4me1/2. Since H3K4me2 is an active histone mark, KDM1 favors gene silencing. The second class of KDMs includes a cluster of six (KDM 2-7) of Fe²⁺/oxoglutarate-dependent enzymes, containing a characteristic Jumonji C (JmjC) domain. Most of the clusters include at least two members and are characterized by one or more targets.

Overexpression or mutation of KDMs has been linked to mis-erased histone methyl modifications and may play oncogenic or tumor-suppressive roles in several neoplasms. Histone demethylases such as KDM5A, 5C, and 6A are involved in several malignancies (acute myeloid leukemia, esophageal, renal, and squamous cell carcinomas) and KDM1A is thought to play a role in several neoplasms including PCa [51]. In PCa, KDM1A can function as AR co-activator and its oncogenic activity is partially due to its ability to trigger Myc-dependent transcription [51], and inhibit p53 pro-apoptotic function [52]. In the presence of high androgen concentrations, KDM1A can be recruited by the AR to mediate gene silencing thus acting as co-repressor [53]. In primary PCa, high KDM1A expression predicts higher risk of tumor relapse after prostatectomy [54]. Other KDMs observed to be overexpressed in CRPC are: KDM4C, found to be required for esophageal squamous carcinomas cell proliferation [55], PHF8 implicated in PCa metastatic disease [56] and KDM5B, an AR-co-activator [57]. Conversely, KDM2A is downregulated in PCa, indicating that it could play a tumor-suppressive function through its role in maintaining genome integrity (Table 20.6) [58]. Moreover, it has been shown that a correlation between KDMs expression and different PCa progression stages indicating that KDMs could have a role as novel biomarkers for prediction of tumor-initiation, progression-free survival, and androgen-independent state [59].

MicroRNAs (miRNAs) are small non-coding RNA, which are approximately 19–22 nucleotides in length. In 1993 and 2000, the first two miRNAs, lin-4 and let-7, were identified in Caenorhabditis elegans [64, 65]. To date, there are 1,872 miRNAs that have been discovered in the human genome (miRBase: Released 20-June 2013). Furthermore, miRNAs have been shown to be involved in the regulation of cell differentiation, development, metabolism, cell cycle, and signaling transduction pathways by targeting 3′ untranslated region (3′ UTR) of target mRNA. This binding leads to degradation of target mRNAs. Dysregulation of miRNAs has been reported in infection diseases, cardiovascular diseases, neurodegenerative diseases, and cancer [66]. For this reason, numerous studies have investigated the function, underlying mechanisms, and therapeutic targeting of miRNAs in multiple disease types. Specific to CRPC, miRNAs have been demonstrated to be involved in disease progression, suggesting their utility as markers for diagnosis/prognosis, as well as novel therapeutic targets (Tables 20.7 and 20.8). MiRNA can also be divided into two categories, oncomir and tumor suppressor miRNA, according to their functions in promoting cell growth or inducing cell death.