**1. Introduction**

Prostate cancer is one of the most frequently diagnosed cancers in Western countries. In advanced prostate cancer, androgen deprivation therapy (ADT) has remained a first-line therapy for decades [1]. After showing an initial response, most patients develop progressive disease, referred to as castration-resistant prostate cancer (CRPC). Intriguingly, CRPC is not androgen independent and several new drugs designed to further suppress the androgen receptor (AR) pathway have led to improved survival, including abiraterone acetate and enzalutamide [2–5]. The human AR gene encodes for a protein with an atomic mass of 110 kDa that consists of an N-terminal domain, a DNA-binding domain, and a ligand-binding domain. AR controls the growth of the prostate gland, and much evidence from preclinical and clinical studies has shown that multiple androgen/AR signaling pathways implicated throughout the various stages of prostate cancer [6]. In addition, recent reports have described the potential therapeutic implications of estrogen and related receptors in prostate cancer [7,8]. However, not all patients respond equally to these newer AR-targeting drugs. Approximately 20–40% of patients with CRPC have a poor clinical response to such agents, and nearly all patients who initially respond acquire secondary resistance. Prospective trials are ongoing to develop the best biomarker strategy for identifying patients resistant to these drugs.

Prostate cancer progresses in a multistep process in response to changes in genetic mechanisms. However, in addition to genetic mutations, epigenetic alterations have also been identified as activating oncogenes and causing a loss of function of tumor suppressor genes [9,10]. Methylation is a form of post-translational covalent modification of histones that epigenetically regulates specific gene expression patterns. Lysine-specific demethylase 1 (LSD1), a member of the flavin adenine dinucleotide dependent enzyme family, behaves like a histone demethylase. LSD1 acts by removing one or two methyl (but not three) groups from lysine residues 4 or 9 in histone H3 (H3K4 and H3K9, respectively) [11].

Growing evidence indicates that LSD1 is critical for human tumorigenesis, and its expression is increased in several malignancies, including prostate, breast, lung, ovarian, and colon cancers [12–16]. Therefore, the inhibition of LSD1 holds promise as a novel anticancer strategy. We have previously developed a novel and selective LSD1 inhibitor called NCL1 ( *N*-[(1S)-3-[3-(trans-2-aminocyclopropyl) phenoxy]-1-(benzylcarbamoyl)propyl] benzamide) after using a combination of in vitro screening and protein structure similarity clustering [17]. In addition, we have reported that NCL1 impairs LSD1 demethylase activity and inhibits cell proliferation in castration-naïve prostate cancer [16].

In this study, we examined the LSD1 status in CRPC cell lines and human specimens including aggressive neuroendocrine differentiated (NED) phenotypes. In addition, we tested the therapeutic significance of NCL1 in CRPC cells in vitro and in an ex vivo subcutaneous model. Furthermore, we investigated the pharmacological mechanism of NCL1 using chromatin immunoprecipitation (ChIP), flow cytometry, and Western blot analyses. To our knowledge, we are the first laboratory to describe the inhibition of LSD1-induced cell death in CRPC through the regulation of autophagy by NCL1.
