MCM2-7 complex is a novel druggable target for neuroendocrine prostate cancer

Neuroendocrine prostate cancer (NEPC) is a lethal subtype of prostate cancer that rarely develops de novo in primary tumors and is commonly acquired during the development of treatment resistance. NEPC is characterized by gain of neuroendocrine markers and loss of androgen receptor (AR), making it resistant to current therapeutic strategies targeting the AR signaling axis. Here, we report that MCM2, MCM3, MCM4, and MCM6 (MCM2/3/4/6) are elevated in human NEPC and high levels of MCM2/3/4/6 are associated with liver metastasis and poor survival in prostate cancer patients. MCM2/3/4/6 are four out of six proteins that form a core DNA helicase (MCM2-7) responsible for unwinding DNA forks during DNA replication. Inhibition of MCM2-7 by treatment with ciprofloxacin inhibits NEPC cell proliferation and migration in vitro, significantly delays NEPC tumor xenograft growth, and partially reverses the neuroendocrine phenotype in vivo. Our study reveals the clinical relevance of MCM2/3/4/6 proteins in NEPC and suggests that inhibition of MCM2-7 may represent a new therapeutic strategy for NEPC.


MCM2/3/4/6 are upregulated in NEPC.
To identify clinically relevant, druggable targets for NEPC, we analyzed our previously reported proteomic analysis of a NEPC model driven by the Trop2 oncogene (TD-NEPC) 16 (Fig. 1A). Four proteins of the MCM2-7 complex, including MCM2/3/4/6, were identified as highly elevated in the TD-NEPC model when compared to LNCaP prostate adenocarcinoma xenografts (Fig. 1A). Elevated gene expression levels of MCM2/3/4/6 were also observed in NCI-H660, a NEPC cell line, when compared to prostate adenocarcinoma cell lines and two CRPC cell lines, 22Rv1 and DU145, showed slight up-regulation of MCM2/3/4/6 genes compared with LNCaP, VCaP, and PC-3 cell lines 17 (Fig. 1B). Moreover, CRPC with either MCM2/3/4/6 gene amplifications or > twofold mRNA up-regulation were associated with worse patient overall survival 18 ( Fig. 1C and Supplementary Fig. 1). Gene expression levels of MCM2/3/4/6 were also specifically elevated in human NEPC when compared to adenocarcinoma CRPC 19,20 ( Fig. 1D-F). Likewise, a positive correlation of mRNA expression levels of MCM2, MCM4 and MCM6 with MCM3, and an inverse correlation with AR downstream targets KLK2 and KLK3 (PSA) was observed 20 (Fig. 1G). Furthermore, relatively higher mRNA levels of MCM2/3/4 were associated with liver metastasis when compared with localized prostate cancer, which may be attributed to the higher incidences of liver metastasis in NEPC when compared to adenocarcinoma CRPC in the SU2C/PCF cohort 18 (Fig. 2). Consistent with our results in human NEPC, we found elevated mRNA levels of MCM2/3/4/6 as well as MCM5 and MCM7 in NEPC PDXs 21 , (Fig. 3A). We performed immunohistochemical (IHC) analysis of PDX tissue microarrays which further confirmed the expression of MCM3, MCM4, MCM5, MCM6, and MCM7 at the protein level in NEPC PDXs ( NEPC is sensitive to MCM inhibition. The MCM2-7 complex is assembled as a double hetero-hexamer of MCM2 to MCM7 (Fig. 5A). The complex functions as a DNA helicase to unwind DNA replication forks at the beginning of DNA replication. Ciprofloxacin is an FDA approved antibiotic previously shown to inhibit   22 . We tested whether inhibition of MCM helicase activity by ciprofloxacin would affect NEPC cell growth. NEPC cells, including TD-NEPC and NCI-H660 cells, showed lower viability in a dose-dependent manner to ciprofloxacin while a prostate adenocarcinoma model (LNCaP) was unaffected (Fig. 5B). We further evaluated cell cycle and apoptosis by measuring DNA content (propidium iodide staining) and cleaved PARP1 by western blot. Ciprofloxacin slowed down the cell cycle at G0/G1 phase, increased the percentage of dead cells in the Sub G1 phase (Fig. 5C) and apoptosis marker, cleaved PARP1 ( Fig. 5D and Supplementary Fig. 4C). Furthermore, treatment with ciprofloxacin significantly delayed the growth of TD-NEPC cells with a high level of MCMs and three CRPC cell lines with a mid-level of MCMs (DU145, PC-3, and 22Rv1) in a dose-dependent manner in colony formation (Fig. 5E). TD-NEPC and 22Rv1 cells were more sensitive to ciprofloxacin when compared to DU145 and PC3 depicted by their growth inhibition at 20 µM ciprofloxacin (Fig. 5E). Ciprofloxacin also inhibited 3D tumorsphere formation and invasion ability of TD-NEPC cells, while DU145 was less sensitive to ciprofloxacin in 3D tumorsphere assay ( Supplementary  Fig. 5). Collectively, inhibition of MCM2-7 DNA-helicase activity with ciprofloxacin dramatically delayed cell proliferation, clonogenicity, and invasion of NEPC cells. Consistent with the heterogenous MCMs pattern in CRPC in PDXs (Fig. 3B, Supplementary Fig. 2 and 3), three CRPC cell lines with mid-level MCMs (DU145, PC-3, and 22Rv1) also responded to ciprofloxacin in colony formation assay.

MCM inhibition delays NEPC tumor growth in vivo and partially reverses the neuroendocrine phenotype.
To test whether inhibition of MCM2-7 helicase activity would affect the growth of NEPC in vivo, we implanted subcutaneously two NEPC xenografts, TD-NEPC and NCI-H660, into immunocompromised male NSG mice. When the average tumor volumes reached 50-75 mm 3 , ciprofloxacin was administered intraperitoneally daily, and the tumor volumes and body weights were measured every 3 days. Ciprofloxacin significantly delayed tumor growth in both TD-NEPC and NCI-H660 NEPC models (Fig. 6A,B). Ciprofloxacin inhibits MCM2-7 activity, but not protein levels and as expected, MCM3 protein levels did not appear to change based on IHC staining. However, neuroendocrine markers (CHGA and SYP) and the proliferation marker (Ki67) showed significantly decreased staining after treatment with ciprofloxacin, while AR expression was not restored (Fig. 6C,D and Supplementary Fig. 6A). At the doses of ciprofloxacin (50 mg/kg) used in this study, toxicity assessed by loss of animal body above 80% was observed at Day 18 of treatment (Supplementary Fig. 6B and 2C). These results indicate that blocking the DNA helicase activity of the MCM2-7 complex using ciprofloxacin significantly delays NEPC tumor growth in vivo and partially reverses the expression of neuroendocrine markers.

Discussion
Currently, there are no long-term effective therapeutic strategies for patients with NEPC. To gain insights into new actionable targets for NEPC, research efforts have been directed to identify key regulators of NEPC development and progression. MYCN and AURKA amplification, RB loss and TP53 mutations, upregulation of BCL2, as well as aberrant expression of transcription factors BRN2, FOXA2, and ONECUT2 have been associated with or implicated in NEPC development [23][24][25][26][27][28][29][30][31][32][33] . Here, we utilized proteomic analysis of a previously characterized NEPC model and identified MCM2, MCM3, MCM4, and MCM6 proteins as novel targets to inhibit NEPC growth. High levels of MCM2/3/4/6 were significantly enriched in human NEPC compared to castration-resistant prostate adenocarcinoma. NEPC can arise from neuroendocrine transdifferentiation of prostate adenocarcinoma and is orchestrated by global epigenetic modifications mediated by SOX2 and EZH2 27,34 . Genistein and trichostatin A (TSA), compounds that have been shown to act through altering epigenetic silencing suppress the expression of all MCM genes in prostate cancer [35][36][37][38][39] . These findings suggest that elevated levels of MCM2/3/4/6 in NEPC may be driven in part by upregulated SOX2 and EZH2 during NEPC development. The precise mechanism underlying the elevated levels of MCM2/3/4/6 and the functional role of MCM2/3/4/6 in NEPC development is yet to be elucidated. www.nature.com/scientificreports/ Ciprofloxacin is an FDA approved antibiotic that is ubiquitously used to treat various bacterial infections and has been shown to inhibit bacterial DNA replication through inhibition of DNA gyrase and DNA topoisomerase. In addition, several studies have demonstrated the ability of ciprofloxacin to induce apoptosis, arrest cell cycle, and inhibit proliferation of human colon, lung, and prostate cancer cells [40][41][42] . In this study, we utilized ciprofloxacin to inhibit MCM2-7 activity in NEPC. We demonstrated that ciprofloxacin significantly delays NEPC cell growth and migration. Furthermore, inhibition of MCM2-7 activity via ciprofloxacin exhibits potent anti-tumor effects in NEPC, reverses neuroendocrine features and reveals a potential new clinically relevant target for NEPC.
Our findings uncover that MCM2/3/4/6 are markedly elevated in patient NEPC and represent new druggable targets for therapeutic intervention. The study further reveals that inhibition of MCM2-7 complex using ciprofloxacin or other targeted approaches may represent a new effective therapy for NEPC. Our studies warrant further exploration of design and screening for MCM2-7 complex-specific inhibitors.

Methods
All methods were carried out in accordance with the relevant guidelines and regulations of Stanford University.

Heatmap, survival and correlation analyses.
Heatmaps of mRNA z-scores were generated using Morpheus (https:// softw are. broad insti tute. org/ morph eus/). Kaplan-Meier survival analysis was performed using Prism 6.0 by comparing samples with no alterations of MCM2/3/4/6 (n = 86) to those with gene amplifications or mRNA upregulation (> twofold) in at least one of the MCMs (MCM2, MCM3, MCM4 or MCM6) (n = 29). Pearson correlation coefficient score of MCM2/3/4/6 and KLK2/KLK3 was acquired from the Fred Hutchinson CRC dataset accessed via cBioPortal Cancer Genomics 43 . A correlation heatmap table was used to visualize the individual association of each gene.
Immunohistochemical staining. Tissue microarrays (TMAs) available at Stanford University comprised of normal prostate tissues and prostate cancers of a spectrum of grades taken from formalin-fixed paraffinembedded radical prostatectomy specimens were used to assess expression of MCM proteins in normal prostate tissues and localized cancer. TMAs of LuCaP patient-derived xenograft models of advanced prostate cancers were constructed from subcutaneous tumors, three tumors per PDX models and three punches per tumor as previously described 21 . TMAs were sectioned at 4 microns and were deparaffinized, rehydrated, and heated to 95 °C for 30 min in 10 mM sodium citrate (pH 6.0) for antigen retrieval. A five-minute incubation in 3% hydrogen peroxide in 1xPBS was used to block endogenous peroxidase activity, and 2.5% horse serum in 1xPBS was applied for 1 h to reduce non-specific background.  www.nature.com/scientificreports/ (sc-393618), anti-MCM7 (sc-9966), anti-AR (sc-7305), anti-SYP (sc-17750), anti-CHGA (sc-393941), and anti-Ki67 (sc-23900) primary antibodies were purchased from Santa Cruz Biotechnology and used at 1:100 dilution. Secondary antibodies were purchased from Vector Labs (MP-7452) and used according to the manufacturer's recommendations. After counterstaining with hematoxylin, the slides were dehydrated, mounted with cover slips, and imaged by Leica DMi8 microscope or Hamamatsu NanoZoomer.
Viability assay. 5000 (TD-NEPC, and LNCaP) or 10,000 (NCI-H660) cancer cells were seeded in 96-well plates and allowed to attach overnight. The following morning, ciprofloxacin was added at the indicated concentrations. After 3 days of treatment, the viable cells were quantified by CellTiter-Blue® Reagent (Promega) and percentage viability was computed by comparison to vehicle control.
Cell cycle analysis by DNA content (propidium iodide staining). 2  Colony formation assay. 500 cells per well were seeded in 6-well plates for indicated number of days dependent on cell lines (TD-NEPC and DU145 for 9 days, PC-3 for 12 days, and 22Rv1 for 15 days). Every 3 days, medium was replaced with fresh medium containing ciprofloxacin (vehicle, 20, 40, and 80 μM). After 9 days, the colonies were fixed and stained with crystal violet. Relative colony formation ability (%) was quantified by measuring colony area per well, and measurements were normalized based on the colony area of vehicle control.
Tumorsphere assay. 500 cells mixed with 50% Matrigel were seeded in 24-well plates for 15 days. Medium containing the indicated concentrations of ciprofloxacin was exchanged every 3 days. Tumorspheres were imaged at day 15 using a Leica stereomicroscope, and quantification was conducted using ImageJ (ImageJ 1.53e; https:// imagej. nih. gov/ ij/ index. html) software by measuring the number of tumorspheres per well based on the RFP reporter signal.
3D Matrigel drop invasion assays. As previously described 16 , a 3D invasion assay was performed in 24-well plates using 5 × 10 4 cancer cells in 10 μl of 100% Matrigel plated as a drop into each well. Imaging was performed on Day 0 and Day 6 using a Celigo Imaging Cytometer (Nexcelom Bioscience). The medium containing vehicle or the indicated concentrations of ciprofloxacin was changed every 3 days. Cell migration area outside of the drop was measured. The relative invasion ability of cancer cells was normalized to the vehicle controls. The orange pseudo-color represents living cells due to RFP fluorescent signals produced by the cell lines.

Effects of ciprofloxacin on xenograft tumor growth in vivo. All animal experimental procedures
were approved by the Institutional Animal Care and Use Committee (IACUC) of Stanford University and in accordance with ARRIVE guidelines. TD-NEPC (1 × 10 6 ) or NCI-H660 (2 × 10 6 ) cancer cells were mixed with 100 μl of Matrigel and implanted subcutaneously into both flanks of male NOD/SCID/IL-2Rγnull (NSG) mice. Before the average tumor volume reached 50-75 mm 3 , mice were randomized to receive either vehicle (0.