- Email: [email protected]
A claudin 3 and claudin 4-targeted Clostridium perfringens protoxin is selectively cytotoxic to PSA-producing prostate cancer cells
Cancer Letters 351 (2014) 260–264
Contents lists available at ScienceDirect
Cancer Letters journal homepage: www.elsevier.com/locate/canlet
A claudin 3 and claudin 4-targeted Clostridium perfringens protoxin is selectively cytotoxic to PSA-producing prostate cancer cells Victor Romanov a, Terry C. Whyard a, Wayne C. Waltzer a, Theodore G. Gabig b,⇑ a b
Department of Urology, Stony Brook University School of Medicine, Stony Brook, NY, United States Department of Medicine, Stony Brook University School of Medicine, Stony Brook, NY, United States
a r t i c l e
i n f o
Article history: Received 19 March 2014 Received in revised form 11 June 2014 Accepted 11 June 2014
Keywords: Prostate cancer Claudins Clostridium perfringens enterotoxin Prostate specific antigen Protoxin
a b s t r a c t Prostate cancer is the second leading cause of non-cutaneous cancer-related death in males, and effective strategies for treatment of metastatic disease are currently limited. The tight junction proteins, claudin 3 and claudin 4, serve as cell-surface receptors for the pore-forming Clostridium perfringens enterotoxin [CPE]. Most prostate cancer cells overexpress claudin 3 and claudin 4, and claudins are aberrantly distributed over the plasma membrane, making these cells particularly sensitive to cytolysis by CPE. Prostate cancer cells secrete PSA locally that is proteolytically active; however, circulating PSA is inactivated via binding to protease inhibitors. To overcome systemic toxicity of CPE, a modified protoxin was constructed with a tethered ligand attached to the C-terminus connected by a flexible linker containing a PSA-specific protease cleavage site. This engineered protoxin selectively and efficiently lyses PSA-producing prostate cancer cells whereas CLDN3 and CLDN4 positive cells that do not express PSA are resistant to cytolysis. Ó 2014 Elsevier Ireland Ltd. All rights reserved.
Introduction Prostate cancer is the most frequently diagnosed male cancer in Western countries [1]. While localized disease is potentially curable with radiation or surgical resection, metastatic disease occurs in 30% of cases, and the mortality rate due to recurrent or initial metastatic disease is approximately 17% [1]. Currently, androgen blockade or ablation as initial therapy for metastatic disease has a high response rate and successfully controls symptoms and disease progression for approximately 1–2 years. However, disease progression inevitably occurs that is unresponsive to current modalities of androgen blockade, a state that is known as castration-resistant prostate cancer [2]. Chemotherapy for castrationresistant prostate cancer with taxotere results in prolongation of mean survival of only 2.9 months [3]. More effective therapy for castration-resistant prostate cancer has been under intensive study and in addition to newer taxanes [3], agents currently in phase III clinical trials or recently FDA-approved are focused on four major areas: novel anti-androgen receptor agents, immunotherapy and immunomodulation, bone microenvironment and anti-angiogenesis, and growth factor receptor inhibition [4,5]. The current work was focused on a new class of targets: the tight junction proteins ⇑ Corresponding author. Address: Stony Brook University School of Medicine, HSC T15-050, Stony Brook, NY 11794-8151, United States. Tel.: +1 631 444 1748; fax: +1 631 444 7530. E-mail address: [email protected] (T.G. Gabig). http://dx.doi.org/10.1016/j.canlet.2014.06.009 0304-3835/Ó 2014 Elsevier Ireland Ltd. All rights reserved.
CLDN3 and CLDN4, cell surface proteins that are over-expressed in prostate cancer [6]. These claudins serve as cellular receptors for the cytolytic pore-forming Clostridium perfringens enterotoxin [CPE], and over-expressing cells are efficiently lysed by CPE in a cell cycle-independent manner [6]. Since systemic administration of CPE is toxic to normal epithelial cells that express CLDN3, CLDN4 [7], as well as CLDN8, CLDN14 [8], and possibly other claudins that contain a conserved asparagine rather than aspartate within the extracellular loop 2 [9], a strategy for eliminating cytotoxicity toward non-prostate epithelial cells is needed, and this problem was addressed in the current study. This is a sequential double-targeting strategy that combines activation of a biologic agent in the vicinity of targeted cells followed by binding of the activated agent to specific cell surface receptors.
Materials and methods Cloning of NH2 terminus His-tagged CPE C. perfringens strain 2917 obtained from American Type Culture Collection (Manassas, VA) was grown from a single colony and used to prepare bacterial DNA with the QIAamp DNA Mini Kit, according to manufacturer’s directions (Qiagen). The bacterial DNA sequence encoding full-length CPE gene was PCR amplified (primer 1, 50 -AGA TGT TAA TCA TAT GAT GCT TAG TAA CAA TTT AAA TCC-30 ; primer 2, 50 -AGG ATC CTT AAA ATT TTT GAA ATA ATA TTG AAT AAG GG-30 ) [6]. The PCR products were digested with the restriction enzymes NdeI/BamHI and cloned into NdeI/BamHI-digested pET16b (Novagen) expression vector to generate an in-frame NH2-terminus, His-tagged CPE expression plasmid, pET 16b-(His) 5-CPE.
261
V. Romanov et al. / Cancer Letters 351 (2014) 260–264 Preparation of CPE-[HSSQKL]-[NPLVPEA] protoxin A synthetic insert containing flexible loop GGSSSGSGSG, PSA cleavage sequence HSSQKL and claudin 3 docking site NPLVPEA [10] was subcloned, in frame, into pET 16b-(His) 5-CPE using BamHI restriction site to C terminal of full size CPE. The sequences of the constructs were confirmed. The complete amino acid sequence of the N-terminus addition is shown in Fig. 1A. The final construct CPE-[HSSQKL][NPLVPEA] hereafter will be denoted CPE-M. Purification of recombinant proteins His-tagged recombinant CPE toxin and CPE-M were prepared from Escherichia coli BL-21 transformed with pET-16b vector (Novagen) encoding CPE or CPE-M. Transformed bacteria were grown overnight at 37 °C, recombinant protein expression was induced for 3 h with 1 mM isopropyl-D-thio-galactoside, the cells were harvested, resuspended in 20 mM Tris–HCl, pH 7.9 binding buffer, and lysed by sonication. The fusion protein was isolated from the supernatant on a His–Bind column (Novagen). After 10-column volumes of binding buffer and six-column volumes of wash buffer, His-tagged CPE was eluted with 0.5 M NaCl, 20 mM Tris–HCl, pH 7.9, and dialyzed (3500 cutoff dialysis tubing) against PBS overnight. Cell culture and cytotoxicity assay The human prostate cancer cell lines LNCaP, PC-3, DU-145 were obtained from ATCC and grown in RPMI-1640 media with 10% fetal bovine serum (FBS) and all necessary additives. C4-2B cells from UroCor Inc. (Oklahoma City, OK) were grown in T-medium (80% Dulbecco’s modified Eagle’s medium (Gibco), 20% F12K (Irving Scientific, Santa Ana, CA), 3 g/l NaHCO3, 100 U/l penicillin G, 100 lg/ml streptomycin, 5 lg/ml insulin, 13.6 pg/ml triiodothyronine, 5 lg/ml transferrin, 0.25 lg/ml biotin, 25 g/ml adenine), supplemented with 5% FBS [11]. RWPE-1 prostate epithelial immortalized cell line was purchased from ATCC and were grown as monolayers in keratinocyte serum free media with supplements as recommended by supplier. Cells were maintained at 37 °C in an atmosphere of air with 5% CO2. Cytotoxicity assay Cells were seeded into 96-well plates at 20,000 (LNCaP) or 5000 (PC-3, DU-145, and C4-2B and RWPE-1) cells per well and cultured for 2 days before treatment. The medium was then replaced with serum-free RPMI-1640 medium containing CPE or CPE-M at 100–500 ng/ml (3–15 nM) with or without PSA inhibitor (borate) or active PSA. Exogenous PSA (EMD Biosciences, San Diego, CA), when used, was added at 5 lg/ml. Cell viability was assayed after 48 h of treatment using CellTiter Blue (Promega, Madison, WI), according to the manufacturer’s instructions. Samples were
measured at Ex 560/Em 590 nm using a spectrophotometer (SpectraMax M5, Molecular Devices, Sunnyvale, CA), and readings of treated cells were normalized to those of untreated cells that were also maintained in serum-free conditions. For each cell line, the experiment was performed three times in replicates of eight.
Gel electrophoresis and immunoblotting Cells were grown on 60-mm tissue culture plates, washed twice with ice-cold PBS, and scraped with 300 ml of ice-cold NaHCO3 buffer [1 mM NaHCO3 and 1 mM phenylmethylsulfonyl fluoride, pH7.5]. They were subsequently collected into a microcentrifuge tube, sonicated for 10 s, and put on ice for 30 min. The cell lysates were then mixed with 2 SDS sample buffer and boiled for 5 min. Total cell lysates were resolved by one-dimensional SDS–polyacrylamide gel electrophoresis (PAGE) and electrophoretically transferred onto a polyvinylidene difluoride membrane (Immobilon; Millipore). The membrane was saturated with PBS containing 4% skim milk and then incubated for 1 h at room temperature with primary antibodies in PBS. Anti-CLDN3 and CLDN4 polyclonal antibodies were purchased from Millipore Inc. (San Diego, CA). After rinsing in PBS containing 0.1% Tween 20, the membrane was incubated for 1 h at room temperature with HRP-conjugated antirabbit or anti-mouse IgG (diluted 1:10,000) in PBS. It was then rinsed again, and finally reacted using an ECL Western blotting detection system (Amersham). To confirm equal protein loading to the gel membrane was stained with Ponceau S before blocking.
Assay of PSA enzymatic activity The hydrolysis of a fluorogenic substrate specific for PSA [12] was monitored on a SpectraMax M5 (Molecular Devices) fluorescence spectrophotometer. Excitation was set at 400 nm and emission at 505 nm. Assays were performed at 25 °C in 50 mM Tris–HCl (pH 7.9) and 100 mM NaCl with Mu-His-Ser-Ser-Lys-Leu-GlnAFC (Calbiochem, La Jolla, CA) as the substrate. Enzymatically active PSA (6 mg Calbiochem) or cell conditioned media was mixed with the fluorogenic substrate dissolved in assay buffer or in assay buffer with added boric acid (10–60 lg/ml Sigma) to a final volume of 600 ll. Increase in fluorescence was recorded for 15 min. Pure AFC (Calbiochem) dissolved in dimethyl sulfoxide at three incremental concentrations was used as a standard to construct a calibration curve. The substrate has some background fluorescence. For this reason, blanks were prepared containing the substrate in assay buffer. In PSA inhibition experiments in the presence of boric acid, readings were consistently lower, probably due to fluorescence quenching. Therefore, boric acid blanks were also prepared. Enzyme activity in conditioned cell culture media and or with boric acid as inhibitor was normalized to exogenous enzymatically active PSA [13].
PSA
(A) S
G Q L
K S S H
S
G
CLDN3 ecl-2
S
NTIIRDFYNPLVPEAQ -KRE-C
N-
290 – - 319 G 1- 199 320 -325 200 - 289 CLDN BINDING MEMBRANE PORE-FORMING
C- Q L K
(B) N-
S
S
G S S S
H
CLDN3 or CLDN4 ecl-2 1- 199 290 – - 319 320 - 325 200 - 289 MEMBRANE PORE-FORMING CLDN BINDING
Fig. 1. Targeted prodrug: proposed mechanism of action. (A) CPE-(HSSKLQ)-(NPLVPEA) construct (CPE-M). CLDN docking site (NPLVPEAQ) blocks the CPE claudin-binding domain. In this conformation binding to CLDN3 or CLDN4 on the cell surface may not be favored. (B) Activation of the protein by PSA cleavage. Following release of the peptide containing the tethered (NPLVPEA) ligand, occupancy of the CPE claudin binding site may favor cell surface CLDN3 or CLDN4. Once activated by PSA cleavage, CPEHSSKLQ may initiate cytolysis via a mechanism similar to that of native CPE.
262
V. Romanov et al. / Cancer Letters 351 (2014) 260–264
siRNA and transfection ON TARGET plus SMART pool siRNA specific for human CLDN3 and CLDN4 and nonspecific control pool were obtained from Thermo Scientific Dharmacon. C4-2B cells were transfected with 20 pmol siRNAs using DharmaFECT 3 transfection reagent (Thermo Scientific Dharmacon) according to the manufacturer’s protocols. Cells were collected 48 h after transfection for expression validation by Western blot and for functional cytotoxicity assays.
Results Comparative cytotoxicity of CPE and protoxin CPE-M For the current study a protoxin consisting of full length CPE with a C-terminus extension containing a CPE binding motif from the extracellular loop 2 (ecl-2) of CLDN3 (NTIIRDFYNPLVPEAQ) [14,15] was engineered. This protoxin construct is hereafter referred to as CPE-M. The two domains of CPE-M were connected by a flexible peptide loop (GGSSSGSGS) followed by a highly specific PSA cleavage site (HSSKLQ) [13] (Fig. 1A). It was hypothesized that the claudin-binding domain of CPE could be blocked by the tethered CLDN3 docking site, and that cleavage at the HSSKLQ site may favor release of the tethered ligand and binding of CPE to the ecl-2 of CLDN3 or CLDN4. A similar strategy was exploited in the design of MMP activated pro-antibodies [16]. The hypothesis was tested by comparing the action of the CPE-M construct with unmodified CPE in cytotoxicity assays against a panel of prostate cancer cells expressing diverse levels of CLDN 3, CLDN4 and enzymatically-active PSA. Four human prostate cancer cell lines and one human normal prostate epithelial line were examined. C4-2B cells produce high levels of CLDN3, CLDN4 and active PSA (Fig. 2A and B). These cells were killed effectively and equally by incubation with either recombinant CPE or CPE-M (Fig. 3A and B). LNCaP cells produce lower amounts of CLDN 3 and CLDN4, and in the absence of androgen activation, low levels of PSA in comparison with C4-2B cells (Fig. 2A and B). Both CPE and CPEM were less toxic against LNCaP cells compared to C4-2B
Fig. 3. (A) Cytotoxicity and cell type specificity of CPE and CPE-M. Cells were incubated with 200 ng/ml of CPE or CPE-M for 1.5 h. Cell viability was tested with CTB assay as explained in the Section ‘Materials and methods’. Results are expressed as per cent of remaining viable cells, mean ± SD, n = 3. Significantly different from non-treated C4-2B cells P < 0.05; significantly different from CPEM treated C4-2B cells, P < 0.05; #significant difference between CPE-M treated DU145 and PC-3 cells, P < 0.05; (B) phase contrast microscopy of C4-2B cells untreated or treated with 200 ng/ml CPE for 1.5 h.
(Fig. 3A). Lower toxicity of CPE against these cells as compared to C4-2B was likely due to the lower CLDN abundance and/or possibly because of the different subcellular location of CLDN 3 and CLDN4 in LNCaP cells. CPE-M induced less cytotoxicity toward LNCaP cells than native CPE (Fig. 3A), likely due to the low amount of PSA proteolytic activity in LNCaP cells (Fig. 2B). PC-3 cells express lower amounts of CLDN3 and CLDN4 compared with all other tested prostate cancer cell lines including DU-145 cells (Fig. 2A). Both DU-145 and PC-3 cell lines express minimal PSA proteolytic activity (Fig. 2B). Recombinant CPE displayed low activity against DU145 cells and was even less active against PC-3 cells, likely due to the less abundant expression of CLDN 3 and 4 in these cells. CPE-M was completely inactive against both cell lines (Fig. 3A). The RWPE-1 immortalized prostate epithelial cell line produced low levels of both CLDNs (Fig. 2A) and a moderate amount of PSA, comparable to that of LNCaP (Fig. 2B). Sensitivity of RWPE-1 to either CPE or CPE-M was also similar to LNCaP (Fig. 3A).
Mechanisms and time course of cytotoxicity
Fig. 2. (A) Western blot of CLDN3 and CLDN4 expression in cell lines used for the study. Ponceau S staining of the membrane with transferred proteins served as loading control. (B) PSA activity in cell lines studied. PSA activity was measured with PSA-specific fluorescent substrate. Background fluorescence detected for VERO cells was subtracted. PSA activity in the conditioned media of C4-2B cells were set for 100%. Results are expressed as per cent as mean ± SD, n = 3. LNCaP, DU-145 and PC-3 cells produced significantly less PSA than C4-2B, P < 0.05.
Following apical exposure to 200 ng/ml CPE, the prostate cancer cell lines used in this study were shown to undergo rapid cell death within 1.5 h that was associated with nearly undetectable caspase 3 and caspase 7 activation as measured by a fluorogenic caspase substrate assay as described before [17]. Additionally, at 1.5 h post-apical exposure of adherent prostate cancer cells to a low concentration of CPE [200 ng/ml], we detected minimal apoptosis as measured by the DAPI staining [data not shown]. The non-apoptotic cell death observed at low concentrations of CPE was also observed in C4-2B cells exposed to 200 ng/ml CPE-M. These results are concordant with published CPE toxicity mechanisms in prostate cancer cell lines [18], and are different from published CPEmediated apoptotic cell death found in intestinal epithelial cell
V. Romanov et al. / Cancer Letters 351 (2014) 260–264
lines [19]. Differences in CPE cytotoxic mechanisms have been extensively studied in vitro and found to depend on a complex number of cellular properties and assay conditions, including CPE concentration, target cell tissue of origin, co-expression of occludin, and apical versus basolateral toxin exposure [19,20]. Reduction of CPE-M cytotoxicity by inhibition of endogenous PSA activity To further elucidate the PSA requirement for cytotoxicity of CPE-M, this protoxin construct was tested against C4-2B cells in the presence or absence of boric acid to inhibit PSA proteolytic activity. Addition of boric acid inhibited CPE-M cytotoxicity in a concentration-dependent manner. Boric acid alone did not induce cell death in C4-2B cells (Fig. 4A). Increased cytotoxicity of CPE-M toward low PSA-producing cells by exogenous PSA For prostate cancer cells that do not produce endogenous enzymatically-active PSA (DU-145 and PC-3) CPE-M did not show substantial cytotoxicity (Fig. 2B). When exogenous active PSA was added to the incubation media, CPE-M became cytotoxic toward DU-145 cells that produce a high level of CLDN3 (Fig. 4B). Addition of enzymatically active PSA alone had no cytotoxic effect on either
263
cell line. With the addition of active PSA, CPE-M displayed increased cytotoxicity toward PC-3 cells (Fig. 4B), which produce a low amount of CLDN3 and a moderate amount of CLDN4 (Fig. 2A). Decreased cytotoxicity of CPE against C4-2b cells after downregulation of CLDN3, 4 expression in these cells To further delineate the role of CLDN3, 4 in the CPE toxicity we performed knock down experiments with siRNA specific for CLDN3, 4. Down regulation of CLDN3 and 4 was confirmed by Western blot for the specific proteins (Fig. 5A). Reduced activity of CPE was observed in C4-2B cells transfected with CLDN3 or 4 specific siRNA but not with scrambled control (Fig. 5B). Discussion Cytotoxicity of C. perfringens enterotoxin has been examined previously in prostate cancer cell lines, in primary metastatic human prostate cancer cells ex vivo [6], and as a locally injected toxin in a mouse model of human prostate cancer [18]. As a prostate cancer therapeutic, CPE represents a novel cell cycleindependent agent with a unique target compared to new agents that are currently in active clinical trials (reviewed in the introduction). However, there are a number of potential obstacles that need to be addressed to demonstrate feasibility of CPE as a
Fig. 4. Effects of modulation of PSA activity on CPE-M cytotoxicity. (A) Cytotoxicity of CPE-M after treatment of PSA-producing cells (C4-2B) with PSA inhibitor-boric acid. C42B cells were treated with or without 200 ng/ml of CPE-M for 1.5 h. BA was added as indicated on the graph. Results are expressed as mean ± SD, n = 3. CPE-M+BA significantly different from CPE-M treated along P < 0.05 and (B) cytotoxicity of CPE-M after treatment PSA nonproducing cells (DU-145, PC-3) with exogenous PSA. CPE-M was added at 200 ng/ml where indicated. Incubation time 1.5 h. Results are expressed as mean ± SD, n = 3. Significantly different from DU-145 cells treated with CPE-M only, P < 0.05; significantly different from PC-3 cells treated with CPE-M only, P < 0.05.
264
V. Romanov et al. / Cancer Letters 351 (2014) 260–264
Financial support This work was support by a Translational Research Opportunity grant from the Stony Brook University Research Foundation. Acknowledgment We greatly thank Dr. Todd Miller (SUNY at Stony Brook) for fruitful discussion and generous advices. References
Fig. 5. Role of CLDN3 and 4 in CPE cytotoxicity. CLDN3, 4 expression was reduced in C4-2b cells by transfecting cells with specific siRNA. (A) Western blot for CLDN3, 4, 48 h after siRNA transfection, b tubulin served as loading control. (B) Reduced cytotoxicity of CPE against CLDN 3 or 4 transfected cells compared to cells transfected with non-specific siRNA, P < 0.05.
therapeutic agent for human prostate cancer, the major one being potential systemic toxicity to normal epithelial tissues that express CLDN3 and CLDN4, including small bowel, colon, pancreas, lung and kidney [6]. The current study was aimed at reduction of non-prostate toxicity, and employed a sequential dual targeting strategy designed to restrict cytotoxicity to PSA-producing target cells. Except for intracellular lysosomal proteases and cathepsin B, the HSSKLQ synthetic PSA recognition site is not cleaved efficiently by other human trypsin- or chymotripsin-like proteases [21]. Thus activation of the protoxin by proteolytic release of the blocking tethered ligand may be restricted to the immediate vicinity of PSA producing cells. The current studies demonstrate proof-of-principle of that off-target cytotoxicity toward PSA-negative, CLDN3 and CLDN4 expressing cells can be greatly reduced or eliminated. Preferential cytotoxicity to prostate cancer cells may be further enhanced in those cancers that over-express CLDN3 and 4 and additionally have lost cell polarity, no longer confining cell surface CLDN to intercellular tight junction complexes and making the CLDN more accessible to CPE-binding [18]. It may thus be possible to achieve clinical target enrichment by selecting those patients whose prostate cancer biopsy specimens show highly over-expressed CLDN3 and 4 with loss of cell polarity as well as high PSA expression. Conflict of Interest We declare no conflicts of interest.
[1] C. Dubois et al., Differential effects of thapsigargin analogues on apoptosis of prostate cancer cells: complex regulation by intracellular calcium, FEBS J. (2013). [2] J.K. Huang et al., Effect of thapsigargin on Ca(2)+ fluxes and viability in human prostate cancer cells, J. Recept. Signal Transduct. Res. 31 (3) (2011) 247–255. [3] K.A. Landers et al., Identification of claudin-4 as a marker highly overexpressed in both primary and metastatic prostate cancer, Br. J. Cancer 99 (3) (2008) 491–501. [4] W. Walther et al., Novel Clostridium perfringens enterotoxin suicide gene therapy for selective treatment of claudin-3- and -4-overexpressing tumors, Gene Ther. 19 (5) (2012) 494–503. [5] J. Meng et al., Testosterone regulates tight junction proteins and influences prostatic autoimmune responses, Horm. Cancer 2 (3) (2011) 145–156. [6] H. Long et al., Expression of Clostridium perfringens enterotoxin receptors claudin-3 and claudin-4 in prostate cancer epithelium, Cancer Res. 61 (21) (2001) 7878–7881. [7] M.E. Fernandez Miyakawa et al., Clostridium perfringens enterotoxin damages the human intestine in vitro, Infect. Immun. 73 (12) (2005) 8407–8410. [8] A. Shrestha, B.A. McClane, Human claudin-8 and -14 are receptors capable of conveying the cytotoxic effects of Clostridium perfringens enterotoxin, MBio 4 (1) (2013). [9] S.L. Robertson, J.G. Smedley 3rd, B.A. McClane, Identification of a claudin-4 residue important for mediating the host cell binding and action of Clostridium perfringens enterotoxin, Infect. Immun. 78 (1) (2010) 505–517. [10] L. Winkler et al., Molecular determinants of the interaction between Clostridium perfringens enterotoxin fragments and claudin-3, J. Biol. Chem. 284 (28) (2009) 18863–18872. [11] T.T. Wu et al., Establishing human prostate cancer cell xenografts in bone: induction of osteoblastic reaction by prostate-specific antigen-producing tumors in athymic and SCID/bg mice using LNCaP and lineage-derived metastatic sublines, Int. J. Cancer 77 (6) (1998) 887–894. [12] K. Turksen, Claudins and cancer stem cells, Stem Cell Rev. 7 (4) (2011) 797–798. [13] M.T. Gallardo-Williams et al., Inhibition of the enzymatic activity of prostatespecific antigen by boric acid and 3-nitrophenyl boronic acid, Prostate 54 (1) (2003) 44–49. [14] Z. Gao, B.A. McClane, Use of Clostridium perfringens enterotoxin and the enterotoxin receptor-binding domain (C-CPE) for cancer treatment: opportunities and challenges, J. Toxicol. 2012 (2012) 981626. [15] S.L. Robertson, B.A. McClane, Interactions between Clostridium perfringens enterotoxin and claudins, Methods Mol. Biol. 762 (2011) 63–75. [16] O. Erster et al., Site-specific targeting of antibody activity in vivo mediated by disease-associated proteases, J. Control. Release 161 (3) (2012) 804–812. [17] V. Romanov et al., Aristolochic acid-induced apoptosis and G2 cell cycle arrest depends on ROS generation and MAP kinases activation, Arch. Toxicol. (2014). [18] T. Maeda et al., Claudin-4-targeted therapy using Clostridium perfringens enterotoxin for prostate cancer, Prostate 72 (4) (2012) 351–360. [19] G. Chakrabarti, X. Zhou, B.A. McClane, Death pathways activated in CaCo-2 cells by Clostridium perfringens enterotoxin, Infect. Immun. 71 (8) (2003) 4260–4270. [20] U. Singh et al., Comparative biochemical and immunocytochemical studies reveal differences in the effects of Clostridium perfringens enterotoxin on polarized CaCo-2 cells versus Vero cells, J. Biol. Chem. 276 (36) (2001) 33402– 33412. [21] S.R. Denmeade et al., Enzymatic activation of a doxorubicin-peptide prodrug by prostate-specific antigen, Cancer Res. 58 (12) (1998) 2537–2540.
Contents lists available at ScienceDirect
Cancer Letters journal homepage: www.elsevier.com/locate/canlet
A claudin 3 and claudin 4-targeted Clostridium perfringens protoxin is selectively cytotoxic to PSA-producing prostate cancer cells Victor Romanov a, Terry C. Whyard a, Wayne C. Waltzer a, Theodore G. Gabig b,⇑ a b
Department of Urology, Stony Brook University School of Medicine, Stony Brook, NY, United States Department of Medicine, Stony Brook University School of Medicine, Stony Brook, NY, United States
a r t i c l e
i n f o
Article history: Received 19 March 2014 Received in revised form 11 June 2014 Accepted 11 June 2014
Keywords: Prostate cancer Claudins Clostridium perfringens enterotoxin Prostate specific antigen Protoxin
a b s t r a c t Prostate cancer is the second leading cause of non-cutaneous cancer-related death in males, and effective strategies for treatment of metastatic disease are currently limited. The tight junction proteins, claudin 3 and claudin 4, serve as cell-surface receptors for the pore-forming Clostridium perfringens enterotoxin [CPE]. Most prostate cancer cells overexpress claudin 3 and claudin 4, and claudins are aberrantly distributed over the plasma membrane, making these cells particularly sensitive to cytolysis by CPE. Prostate cancer cells secrete PSA locally that is proteolytically active; however, circulating PSA is inactivated via binding to protease inhibitors. To overcome systemic toxicity of CPE, a modified protoxin was constructed with a tethered ligand attached to the C-terminus connected by a flexible linker containing a PSA-specific protease cleavage site. This engineered protoxin selectively and efficiently lyses PSA-producing prostate cancer cells whereas CLDN3 and CLDN4 positive cells that do not express PSA are resistant to cytolysis. Ó 2014 Elsevier Ireland Ltd. All rights reserved.
Introduction Prostate cancer is the most frequently diagnosed male cancer in Western countries [1]. While localized disease is potentially curable with radiation or surgical resection, metastatic disease occurs in 30% of cases, and the mortality rate due to recurrent or initial metastatic disease is approximately 17% [1]. Currently, androgen blockade or ablation as initial therapy for metastatic disease has a high response rate and successfully controls symptoms and disease progression for approximately 1–2 years. However, disease progression inevitably occurs that is unresponsive to current modalities of androgen blockade, a state that is known as castration-resistant prostate cancer [2]. Chemotherapy for castrationresistant prostate cancer with taxotere results in prolongation of mean survival of only 2.9 months [3]. More effective therapy for castration-resistant prostate cancer has been under intensive study and in addition to newer taxanes [3], agents currently in phase III clinical trials or recently FDA-approved are focused on four major areas: novel anti-androgen receptor agents, immunotherapy and immunomodulation, bone microenvironment and anti-angiogenesis, and growth factor receptor inhibition [4,5]. The current work was focused on a new class of targets: the tight junction proteins ⇑ Corresponding author. Address: Stony Brook University School of Medicine, HSC T15-050, Stony Brook, NY 11794-8151, United States. Tel.: +1 631 444 1748; fax: +1 631 444 7530. E-mail address: [email protected] (T.G. Gabig). http://dx.doi.org/10.1016/j.canlet.2014.06.009 0304-3835/Ó 2014 Elsevier Ireland Ltd. All rights reserved.
CLDN3 and CLDN4, cell surface proteins that are over-expressed in prostate cancer [6]. These claudins serve as cellular receptors for the cytolytic pore-forming Clostridium perfringens enterotoxin [CPE], and over-expressing cells are efficiently lysed by CPE in a cell cycle-independent manner [6]. Since systemic administration of CPE is toxic to normal epithelial cells that express CLDN3, CLDN4 [7], as well as CLDN8, CLDN14 [8], and possibly other claudins that contain a conserved asparagine rather than aspartate within the extracellular loop 2 [9], a strategy for eliminating cytotoxicity toward non-prostate epithelial cells is needed, and this problem was addressed in the current study. This is a sequential double-targeting strategy that combines activation of a biologic agent in the vicinity of targeted cells followed by binding of the activated agent to specific cell surface receptors.
Materials and methods Cloning of NH2 terminus His-tagged CPE C. perfringens strain 2917 obtained from American Type Culture Collection (Manassas, VA) was grown from a single colony and used to prepare bacterial DNA with the QIAamp DNA Mini Kit, according to manufacturer’s directions (Qiagen). The bacterial DNA sequence encoding full-length CPE gene was PCR amplified (primer 1, 50 -AGA TGT TAA TCA TAT GAT GCT TAG TAA CAA TTT AAA TCC-30 ; primer 2, 50 -AGG ATC CTT AAA ATT TTT GAA ATA ATA TTG AAT AAG GG-30 ) [6]. The PCR products were digested with the restriction enzymes NdeI/BamHI and cloned into NdeI/BamHI-digested pET16b (Novagen) expression vector to generate an in-frame NH2-terminus, His-tagged CPE expression plasmid, pET 16b-(His) 5-CPE.
261
V. Romanov et al. / Cancer Letters 351 (2014) 260–264 Preparation of CPE-[HSSQKL]-[NPLVPEA] protoxin A synthetic insert containing flexible loop GGSSSGSGSG, PSA cleavage sequence HSSQKL and claudin 3 docking site NPLVPEA [10] was subcloned, in frame, into pET 16b-(His) 5-CPE using BamHI restriction site to C terminal of full size CPE. The sequences of the constructs were confirmed. The complete amino acid sequence of the N-terminus addition is shown in Fig. 1A. The final construct CPE-[HSSQKL][NPLVPEA] hereafter will be denoted CPE-M. Purification of recombinant proteins His-tagged recombinant CPE toxin and CPE-M were prepared from Escherichia coli BL-21 transformed with pET-16b vector (Novagen) encoding CPE or CPE-M. Transformed bacteria were grown overnight at 37 °C, recombinant protein expression was induced for 3 h with 1 mM isopropyl-D-thio-galactoside, the cells were harvested, resuspended in 20 mM Tris–HCl, pH 7.9 binding buffer, and lysed by sonication. The fusion protein was isolated from the supernatant on a His–Bind column (Novagen). After 10-column volumes of binding buffer and six-column volumes of wash buffer, His-tagged CPE was eluted with 0.5 M NaCl, 20 mM Tris–HCl, pH 7.9, and dialyzed (3500 cutoff dialysis tubing) against PBS overnight. Cell culture and cytotoxicity assay The human prostate cancer cell lines LNCaP, PC-3, DU-145 were obtained from ATCC and grown in RPMI-1640 media with 10% fetal bovine serum (FBS) and all necessary additives. C4-2B cells from UroCor Inc. (Oklahoma City, OK) were grown in T-medium (80% Dulbecco’s modified Eagle’s medium (Gibco), 20% F12K (Irving Scientific, Santa Ana, CA), 3 g/l NaHCO3, 100 U/l penicillin G, 100 lg/ml streptomycin, 5 lg/ml insulin, 13.6 pg/ml triiodothyronine, 5 lg/ml transferrin, 0.25 lg/ml biotin, 25 g/ml adenine), supplemented with 5% FBS [11]. RWPE-1 prostate epithelial immortalized cell line was purchased from ATCC and were grown as monolayers in keratinocyte serum free media with supplements as recommended by supplier. Cells were maintained at 37 °C in an atmosphere of air with 5% CO2. Cytotoxicity assay Cells were seeded into 96-well plates at 20,000 (LNCaP) or 5000 (PC-3, DU-145, and C4-2B and RWPE-1) cells per well and cultured for 2 days before treatment. The medium was then replaced with serum-free RPMI-1640 medium containing CPE or CPE-M at 100–500 ng/ml (3–15 nM) with or without PSA inhibitor (borate) or active PSA. Exogenous PSA (EMD Biosciences, San Diego, CA), when used, was added at 5 lg/ml. Cell viability was assayed after 48 h of treatment using CellTiter Blue (Promega, Madison, WI), according to the manufacturer’s instructions. Samples were
measured at Ex 560/Em 590 nm using a spectrophotometer (SpectraMax M5, Molecular Devices, Sunnyvale, CA), and readings of treated cells were normalized to those of untreated cells that were also maintained in serum-free conditions. For each cell line, the experiment was performed three times in replicates of eight.
Gel electrophoresis and immunoblotting Cells were grown on 60-mm tissue culture plates, washed twice with ice-cold PBS, and scraped with 300 ml of ice-cold NaHCO3 buffer [1 mM NaHCO3 and 1 mM phenylmethylsulfonyl fluoride, pH7.5]. They were subsequently collected into a microcentrifuge tube, sonicated for 10 s, and put on ice for 30 min. The cell lysates were then mixed with 2 SDS sample buffer and boiled for 5 min. Total cell lysates were resolved by one-dimensional SDS–polyacrylamide gel electrophoresis (PAGE) and electrophoretically transferred onto a polyvinylidene difluoride membrane (Immobilon; Millipore). The membrane was saturated with PBS containing 4% skim milk and then incubated for 1 h at room temperature with primary antibodies in PBS. Anti-CLDN3 and CLDN4 polyclonal antibodies were purchased from Millipore Inc. (San Diego, CA). After rinsing in PBS containing 0.1% Tween 20, the membrane was incubated for 1 h at room temperature with HRP-conjugated antirabbit or anti-mouse IgG (diluted 1:10,000) in PBS. It was then rinsed again, and finally reacted using an ECL Western blotting detection system (Amersham). To confirm equal protein loading to the gel membrane was stained with Ponceau S before blocking.
Assay of PSA enzymatic activity The hydrolysis of a fluorogenic substrate specific for PSA [12] was monitored on a SpectraMax M5 (Molecular Devices) fluorescence spectrophotometer. Excitation was set at 400 nm and emission at 505 nm. Assays were performed at 25 °C in 50 mM Tris–HCl (pH 7.9) and 100 mM NaCl with Mu-His-Ser-Ser-Lys-Leu-GlnAFC (Calbiochem, La Jolla, CA) as the substrate. Enzymatically active PSA (6 mg Calbiochem) or cell conditioned media was mixed with the fluorogenic substrate dissolved in assay buffer or in assay buffer with added boric acid (10–60 lg/ml Sigma) to a final volume of 600 ll. Increase in fluorescence was recorded for 15 min. Pure AFC (Calbiochem) dissolved in dimethyl sulfoxide at three incremental concentrations was used as a standard to construct a calibration curve. The substrate has some background fluorescence. For this reason, blanks were prepared containing the substrate in assay buffer. In PSA inhibition experiments in the presence of boric acid, readings were consistently lower, probably due to fluorescence quenching. Therefore, boric acid blanks were also prepared. Enzyme activity in conditioned cell culture media and or with boric acid as inhibitor was normalized to exogenous enzymatically active PSA [13].
PSA
(A) S
G Q L
K S S H
S
G
CLDN3 ecl-2
S
NTIIRDFYNPLVPEAQ -KRE-C
N-
290 – - 319 G 1- 199 320 -325 200 - 289 CLDN BINDING MEMBRANE PORE-FORMING
C- Q L K
(B) N-
S
S
G S S S
H
CLDN3 or CLDN4 ecl-2 1- 199 290 – - 319 320 - 325 200 - 289 MEMBRANE PORE-FORMING CLDN BINDING
Fig. 1. Targeted prodrug: proposed mechanism of action. (A) CPE-(HSSKLQ)-(NPLVPEA) construct (CPE-M). CLDN docking site (NPLVPEAQ) blocks the CPE claudin-binding domain. In this conformation binding to CLDN3 or CLDN4 on the cell surface may not be favored. (B) Activation of the protein by PSA cleavage. Following release of the peptide containing the tethered (NPLVPEA) ligand, occupancy of the CPE claudin binding site may favor cell surface CLDN3 or CLDN4. Once activated by PSA cleavage, CPEHSSKLQ may initiate cytolysis via a mechanism similar to that of native CPE.
262
V. Romanov et al. / Cancer Letters 351 (2014) 260–264
siRNA and transfection ON TARGET plus SMART pool siRNA specific for human CLDN3 and CLDN4 and nonspecific control pool were obtained from Thermo Scientific Dharmacon. C4-2B cells were transfected with 20 pmol siRNAs using DharmaFECT 3 transfection reagent (Thermo Scientific Dharmacon) according to the manufacturer’s protocols. Cells were collected 48 h after transfection for expression validation by Western blot and for functional cytotoxicity assays.
Results Comparative cytotoxicity of CPE and protoxin CPE-M For the current study a protoxin consisting of full length CPE with a C-terminus extension containing a CPE binding motif from the extracellular loop 2 (ecl-2) of CLDN3 (NTIIRDFYNPLVPEAQ) [14,15] was engineered. This protoxin construct is hereafter referred to as CPE-M. The two domains of CPE-M were connected by a flexible peptide loop (GGSSSGSGS) followed by a highly specific PSA cleavage site (HSSKLQ) [13] (Fig. 1A). It was hypothesized that the claudin-binding domain of CPE could be blocked by the tethered CLDN3 docking site, and that cleavage at the HSSKLQ site may favor release of the tethered ligand and binding of CPE to the ecl-2 of CLDN3 or CLDN4. A similar strategy was exploited in the design of MMP activated pro-antibodies [16]. The hypothesis was tested by comparing the action of the CPE-M construct with unmodified CPE in cytotoxicity assays against a panel of prostate cancer cells expressing diverse levels of CLDN 3, CLDN4 and enzymatically-active PSA. Four human prostate cancer cell lines and one human normal prostate epithelial line were examined. C4-2B cells produce high levels of CLDN3, CLDN4 and active PSA (Fig. 2A and B). These cells were killed effectively and equally by incubation with either recombinant CPE or CPE-M (Fig. 3A and B). LNCaP cells produce lower amounts of CLDN 3 and CLDN4, and in the absence of androgen activation, low levels of PSA in comparison with C4-2B cells (Fig. 2A and B). Both CPE and CPEM were less toxic against LNCaP cells compared to C4-2B
Fig. 3. (A) Cytotoxicity and cell type specificity of CPE and CPE-M. Cells were incubated with 200 ng/ml of CPE or CPE-M for 1.5 h. Cell viability was tested with CTB assay as explained in the Section ‘Materials and methods’. Results are expressed as per cent of remaining viable cells, mean ± SD, n = 3. Significantly different from non-treated C4-2B cells P < 0.05; significantly different from CPEM treated C4-2B cells, P < 0.05; #significant difference between CPE-M treated DU145 and PC-3 cells, P < 0.05; (B) phase contrast microscopy of C4-2B cells untreated or treated with 200 ng/ml CPE for 1.5 h.
(Fig. 3A). Lower toxicity of CPE against these cells as compared to C4-2B was likely due to the lower CLDN abundance and/or possibly because of the different subcellular location of CLDN 3 and CLDN4 in LNCaP cells. CPE-M induced less cytotoxicity toward LNCaP cells than native CPE (Fig. 3A), likely due to the low amount of PSA proteolytic activity in LNCaP cells (Fig. 2B). PC-3 cells express lower amounts of CLDN3 and CLDN4 compared with all other tested prostate cancer cell lines including DU-145 cells (Fig. 2A). Both DU-145 and PC-3 cell lines express minimal PSA proteolytic activity (Fig. 2B). Recombinant CPE displayed low activity against DU145 cells and was even less active against PC-3 cells, likely due to the less abundant expression of CLDN 3 and 4 in these cells. CPE-M was completely inactive against both cell lines (Fig. 3A). The RWPE-1 immortalized prostate epithelial cell line produced low levels of both CLDNs (Fig. 2A) and a moderate amount of PSA, comparable to that of LNCaP (Fig. 2B). Sensitivity of RWPE-1 to either CPE or CPE-M was also similar to LNCaP (Fig. 3A).
Mechanisms and time course of cytotoxicity
Fig. 2. (A) Western blot of CLDN3 and CLDN4 expression in cell lines used for the study. Ponceau S staining of the membrane with transferred proteins served as loading control. (B) PSA activity in cell lines studied. PSA activity was measured with PSA-specific fluorescent substrate. Background fluorescence detected for VERO cells was subtracted. PSA activity in the conditioned media of C4-2B cells were set for 100%. Results are expressed as per cent as mean ± SD, n = 3. LNCaP, DU-145 and PC-3 cells produced significantly less PSA than C4-2B, P < 0.05.
Following apical exposure to 200 ng/ml CPE, the prostate cancer cell lines used in this study were shown to undergo rapid cell death within 1.5 h that was associated with nearly undetectable caspase 3 and caspase 7 activation as measured by a fluorogenic caspase substrate assay as described before [17]. Additionally, at 1.5 h post-apical exposure of adherent prostate cancer cells to a low concentration of CPE [200 ng/ml], we detected minimal apoptosis as measured by the DAPI staining [data not shown]. The non-apoptotic cell death observed at low concentrations of CPE was also observed in C4-2B cells exposed to 200 ng/ml CPE-M. These results are concordant with published CPE toxicity mechanisms in prostate cancer cell lines [18], and are different from published CPEmediated apoptotic cell death found in intestinal epithelial cell
V. Romanov et al. / Cancer Letters 351 (2014) 260–264
lines [19]. Differences in CPE cytotoxic mechanisms have been extensively studied in vitro and found to depend on a complex number of cellular properties and assay conditions, including CPE concentration, target cell tissue of origin, co-expression of occludin, and apical versus basolateral toxin exposure [19,20]. Reduction of CPE-M cytotoxicity by inhibition of endogenous PSA activity To further elucidate the PSA requirement for cytotoxicity of CPE-M, this protoxin construct was tested against C4-2B cells in the presence or absence of boric acid to inhibit PSA proteolytic activity. Addition of boric acid inhibited CPE-M cytotoxicity in a concentration-dependent manner. Boric acid alone did not induce cell death in C4-2B cells (Fig. 4A). Increased cytotoxicity of CPE-M toward low PSA-producing cells by exogenous PSA For prostate cancer cells that do not produce endogenous enzymatically-active PSA (DU-145 and PC-3) CPE-M did not show substantial cytotoxicity (Fig. 2B). When exogenous active PSA was added to the incubation media, CPE-M became cytotoxic toward DU-145 cells that produce a high level of CLDN3 (Fig. 4B). Addition of enzymatically active PSA alone had no cytotoxic effect on either
263
cell line. With the addition of active PSA, CPE-M displayed increased cytotoxicity toward PC-3 cells (Fig. 4B), which produce a low amount of CLDN3 and a moderate amount of CLDN4 (Fig. 2A). Decreased cytotoxicity of CPE against C4-2b cells after downregulation of CLDN3, 4 expression in these cells To further delineate the role of CLDN3, 4 in the CPE toxicity we performed knock down experiments with siRNA specific for CLDN3, 4. Down regulation of CLDN3 and 4 was confirmed by Western blot for the specific proteins (Fig. 5A). Reduced activity of CPE was observed in C4-2B cells transfected with CLDN3 or 4 specific siRNA but not with scrambled control (Fig. 5B). Discussion Cytotoxicity of C. perfringens enterotoxin has been examined previously in prostate cancer cell lines, in primary metastatic human prostate cancer cells ex vivo [6], and as a locally injected toxin in a mouse model of human prostate cancer [18]. As a prostate cancer therapeutic, CPE represents a novel cell cycleindependent agent with a unique target compared to new agents that are currently in active clinical trials (reviewed in the introduction). However, there are a number of potential obstacles that need to be addressed to demonstrate feasibility of CPE as a
Fig. 4. Effects of modulation of PSA activity on CPE-M cytotoxicity. (A) Cytotoxicity of CPE-M after treatment of PSA-producing cells (C4-2B) with PSA inhibitor-boric acid. C42B cells were treated with or without 200 ng/ml of CPE-M for 1.5 h. BA was added as indicated on the graph. Results are expressed as mean ± SD, n = 3. CPE-M+BA significantly different from CPE-M treated along P < 0.05 and (B) cytotoxicity of CPE-M after treatment PSA nonproducing cells (DU-145, PC-3) with exogenous PSA. CPE-M was added at 200 ng/ml where indicated. Incubation time 1.5 h. Results are expressed as mean ± SD, n = 3. Significantly different from DU-145 cells treated with CPE-M only, P < 0.05; significantly different from PC-3 cells treated with CPE-M only, P < 0.05.
264
V. Romanov et al. / Cancer Letters 351 (2014) 260–264
Financial support This work was support by a Translational Research Opportunity grant from the Stony Brook University Research Foundation. Acknowledgment We greatly thank Dr. Todd Miller (SUNY at Stony Brook) for fruitful discussion and generous advices. References
Fig. 5. Role of CLDN3 and 4 in CPE cytotoxicity. CLDN3, 4 expression was reduced in C4-2b cells by transfecting cells with specific siRNA. (A) Western blot for CLDN3, 4, 48 h after siRNA transfection, b tubulin served as loading control. (B) Reduced cytotoxicity of CPE against CLDN 3 or 4 transfected cells compared to cells transfected with non-specific siRNA, P < 0.05.
therapeutic agent for human prostate cancer, the major one being potential systemic toxicity to normal epithelial tissues that express CLDN3 and CLDN4, including small bowel, colon, pancreas, lung and kidney [6]. The current study was aimed at reduction of non-prostate toxicity, and employed a sequential dual targeting strategy designed to restrict cytotoxicity to PSA-producing target cells. Except for intracellular lysosomal proteases and cathepsin B, the HSSKLQ synthetic PSA recognition site is not cleaved efficiently by other human trypsin- or chymotripsin-like proteases [21]. Thus activation of the protoxin by proteolytic release of the blocking tethered ligand may be restricted to the immediate vicinity of PSA producing cells. The current studies demonstrate proof-of-principle of that off-target cytotoxicity toward PSA-negative, CLDN3 and CLDN4 expressing cells can be greatly reduced or eliminated. Preferential cytotoxicity to prostate cancer cells may be further enhanced in those cancers that over-express CLDN3 and 4 and additionally have lost cell polarity, no longer confining cell surface CLDN to intercellular tight junction complexes and making the CLDN more accessible to CPE-binding [18]. It may thus be possible to achieve clinical target enrichment by selecting those patients whose prostate cancer biopsy specimens show highly over-expressed CLDN3 and 4 with loss of cell polarity as well as high PSA expression. Conflict of Interest We declare no conflicts of interest.
[1] C. Dubois et al., Differential effects of thapsigargin analogues on apoptosis of prostate cancer cells: complex regulation by intracellular calcium, FEBS J. (2013). [2] J.K. Huang et al., Effect of thapsigargin on Ca(2)+ fluxes and viability in human prostate cancer cells, J. Recept. Signal Transduct. Res. 31 (3) (2011) 247–255. [3] K.A. Landers et al., Identification of claudin-4 as a marker highly overexpressed in both primary and metastatic prostate cancer, Br. J. Cancer 99 (3) (2008) 491–501. [4] W. Walther et al., Novel Clostridium perfringens enterotoxin suicide gene therapy for selective treatment of claudin-3- and -4-overexpressing tumors, Gene Ther. 19 (5) (2012) 494–503. [5] J. Meng et al., Testosterone regulates tight junction proteins and influences prostatic autoimmune responses, Horm. Cancer 2 (3) (2011) 145–156. [6] H. Long et al., Expression of Clostridium perfringens enterotoxin receptors claudin-3 and claudin-4 in prostate cancer epithelium, Cancer Res. 61 (21) (2001) 7878–7881. [7] M.E. Fernandez Miyakawa et al., Clostridium perfringens enterotoxin damages the human intestine in vitro, Infect. Immun. 73 (12) (2005) 8407–8410. [8] A. Shrestha, B.A. McClane, Human claudin-8 and -14 are receptors capable of conveying the cytotoxic effects of Clostridium perfringens enterotoxin, MBio 4 (1) (2013). [9] S.L. Robertson, J.G. Smedley 3rd, B.A. McClane, Identification of a claudin-4 residue important for mediating the host cell binding and action of Clostridium perfringens enterotoxin, Infect. Immun. 78 (1) (2010) 505–517. [10] L. Winkler et al., Molecular determinants of the interaction between Clostridium perfringens enterotoxin fragments and claudin-3, J. Biol. Chem. 284 (28) (2009) 18863–18872. [11] T.T. Wu et al., Establishing human prostate cancer cell xenografts in bone: induction of osteoblastic reaction by prostate-specific antigen-producing tumors in athymic and SCID/bg mice using LNCaP and lineage-derived metastatic sublines, Int. J. Cancer 77 (6) (1998) 887–894. [12] K. Turksen, Claudins and cancer stem cells, Stem Cell Rev. 7 (4) (2011) 797–798. [13] M.T. Gallardo-Williams et al., Inhibition of the enzymatic activity of prostatespecific antigen by boric acid and 3-nitrophenyl boronic acid, Prostate 54 (1) (2003) 44–49. [14] Z. Gao, B.A. McClane, Use of Clostridium perfringens enterotoxin and the enterotoxin receptor-binding domain (C-CPE) for cancer treatment: opportunities and challenges, J. Toxicol. 2012 (2012) 981626. [15] S.L. Robertson, B.A. McClane, Interactions between Clostridium perfringens enterotoxin and claudins, Methods Mol. Biol. 762 (2011) 63–75. [16] O. Erster et al., Site-specific targeting of antibody activity in vivo mediated by disease-associated proteases, J. Control. Release 161 (3) (2012) 804–812. [17] V. Romanov et al., Aristolochic acid-induced apoptosis and G2 cell cycle arrest depends on ROS generation and MAP kinases activation, Arch. Toxicol. (2014). [18] T. Maeda et al., Claudin-4-targeted therapy using Clostridium perfringens enterotoxin for prostate cancer, Prostate 72 (4) (2012) 351–360. [19] G. Chakrabarti, X. Zhou, B.A. McClane, Death pathways activated in CaCo-2 cells by Clostridium perfringens enterotoxin, Infect. Immun. 71 (8) (2003) 4260–4270. [20] U. Singh et al., Comparative biochemical and immunocytochemical studies reveal differences in the effects of Clostridium perfringens enterotoxin on polarized CaCo-2 cells versus Vero cells, J. Biol. Chem. 276 (36) (2001) 33402– 33412. [21] S.R. Denmeade et al., Enzymatic activation of a doxorubicin-peptide prodrug by prostate-specific antigen, Cancer Res. 58 (12) (1998) 2537–2540.