Intracellular Fas Ligand Expression Causes Fas-Mediated Apoptosis in Human Prostate Cancer Cells Resistant to Monoclonal Antibody-Induced Apoptosis

ARTICLE

doi:10.1006/mthe.2000.0139, available online at http://www.idealibrary.com on IDEAL

Intracellular Fas Ligand Expression Causes Fas-Mediated Apoptosis in Human Prostate Cancer Cells Resistant to Monoclonal Antibody-Induced Apoptosis Marc L. Hyer, Christina Voelkel-Johnson, Semyon Rubinchik, Jian-yun Dong, and James S. Norris1 Department of Microbiology and Immunology, Medical University of South Carolina, Charleston, South Carolina 29425 Received for publication June 8, 2000, and accepted in revised form August 29, 2000

Several laboratories have attempted with little success to induce Fas-mediated apoptosis in prostate cancer (PCa) cells, using different external Fas agonists, i.e., anti-Fas antibodies and membrane-bound FasL. The present study confirms these earlier results using the anti-Fas antibody CH-11 in five human PCa cell lines (PPC-1, LNCaP, PC-3, TSU-Pr1, and DU145). However, intracellular murine FasL expression induced Fas-mediated apoptosis in all CH-11-resistant cell lines. Adenovirus (AdGFPFasLTET) was used to deliver a Murine FasL–GFP fusion gene into human PCa cells resulting in 70–98% apoptosis at 48 h as determined by the MTS assay. DU145 and PPC1 cells treated with AdGFPFasLTET stained positive for the TUNEL assay, indicating that cell death was via apoptosis. Using immunofluorescent microscopy, Fas and GFPFasL colocalized to the same intracellular compartment. The anti-Fas neutralizing antibody ZB-4 was unable to block AdGFPFasLTET-mediated cell death, suggesting that intracellular FasL may ligate Fas within the Golgi and/or endoplasmic reticulum. This is the first evidence suggesting that these two molecules interact prior to cell surface presentation. Collectively, these findings indicate that intracellular GFPFasL expression is superior to CH-11 at inducing Fas-mediated apoptosis in human PCa cells and may allow use of AdGFPFasLTET for PCa gene therapy. Key Words: adenovirus; gene therapy; fas ligand; prostate; apoptosis.

INTRODUCTION Prostate cancer (PCa)2 is the most common noncutaneous malignant cancer diagnosed in humans (1). It is estimated that 179,300 new PCa cases will have been detected in 1999 in the United States alone (1). Current treatment options for localized PCa include radical prostatectomy, cryotherapy, brachytherapy, and external beam radiation. These treatment options have proven benefits and can lead to a cure for patients presenting with localized disease. Unfortunately, many men do not present with PCa until the cancer has become locally advanced and/or metastatic. At this stage the only treatment options avail-

1To whom correspondence and reprint requests should be addressed. Fax: (843) 792-2464. E-mail: [email protected]. 2Abbreviations used: PCa, prostate cancer; m.o.i., multiplicity of infection; CDDP, cis-diamminedichloroplatinum(II); RCA, replicationcompetent adenovirus; VP-16, etoposide; ADR, adriamycin; Ad5, adenovirus serotype 5; ITR, inverted terminal repeat; TRE, tetracycline response element; tTA, tetracycline-controlled transactivator; 7-AAD, 7amino-actinomycin D; ER, endoplasmic reticulum; DISC, death-inducing signaling complex.

348

able are chemotherapy, total androgen ablation, or watchful waiting. Chemotherapy has not become a viable option for treating metastatic PCa because tumors can be slow growing. Androgen ablation prolongs survival on average only 18 months since tumors normally progress to an androgen independent state (2). The lack of effective treatments for locally advanced or metastatic PCa emphasizes the need to develop innovative treatment options such as gene therapy. Novel experimental approaches for the treatment of PCa include delivery of oncolytic viruses (3), immunomodulatory molecules, p53 and p21, enzymes to metabolize prodrugs, and agents that induce apoptosis (4). FasL (CD95L or APO-1L) is a 40-kDa type II membrane protein belonging to the tumor necrosis factor (TNF) family. Its receptor, Fas (CD95 or APO-1) is a 45-kDa type I membrane protein belonging to the TNF/NGF (nerve growth factor) superfamily of receptors (5, 6). Interaction of Fas with its ligand can result in apoptosis of Fas bearing cells. It is believed that ligation of Fas leads to receptor trimerization allowing assembly of a death-inducing signaling complex (DISC) involving association of the death MOLECULAR THERAPY Vol. 2, No. 4, October 2000 Copyright  The American Society of Gene Therapy 1525-0016/00 $35.00

ARTICLE domains (DD) of Fas and one of its associated proteins, FADD/MORT1 (7, 8). Procaspase 8 (FLICE/MACH) is recruited to the signaling complex, where it associates with FADD (9, 10), and becomes autocatalytically activated (8). Active caspase 8 is the initiator caspase that cleaves downstream pro-effector caspases. Activated effector caspases, i.e., caspase 3, are responsible for degrading vital cellular substrates which ultimately leads to apoptosis of the cell. Fas is ubiquitously expressed in most tissue types including prostate tissue (11) and cultured PCa cell lines (12–14). In contrast, FasL is predominantly expressed in immune privileged tissues such as testis (12), retina (15, 16), cornea (17), and placenta (12). More recently, several studies have reported FasL expression in cultured PCa cells (13) as well as mouse (14, 18) and human prostate tissue (12). Suzuki et al. demonstrated in mice that a functional Fas/FasL pathway exists within the prostate (19). In their report, castration of normal mice resulted in prostate gland regression. In contrast, castration of the lpr mouse (20), which lacks a functional Fas receptor due to a mutation in the Fas gene, resulted in no change in prostate gland size (19). A similar mechanism may be involved in atrophy of the human prostate gland caused by castration or hormonal therapy although this has yet to be demonstrated in vivo. It has been shown in vitro that a functional Fas-mediated pathway exists in certain PCa cell lines when coincubated with FasL expressing lymphocytes (21). To take advantage of this existing pathway, we and others (22) have developed FasL-expressing adenoviruses that may be useful for a gene therapy approach to treat advanced PCa. Hedlund et al. generated an adenovirus in which murine FasL was driven by CMV and demonstrated that this virus was able to induce apoptosis in PCa cell lines. However, effectiveness of Ad-FasL to induce cell death in LNCaP and DU-145 cells, which are resistant to Fas antibody induced apoptosis, was inconclusive. In this study, we have generated an adenoviral construct in which murine FasL is fused to the GFP reporter. This construct is able to overcome resistance to Fas-antibodymediated apoptosis in all PCa cell lines tested, including LNCaP, DU145, PC-3, TSU-Pr1, and PPC-1, and the GFP moiety allows us to localize FasL within the cell. Using immunocytochemistry to localize Fas receptor, we colocalized GFPFasL and Fas to the same intracellular compartment, likely the Golgi (23) and/or endoplasmic reticulum. We were unable to block AdGFPFasLTET induced apoptosis with an anti-Fas neutralizing antibody (ZB-4). This is the first evidence indicating these two molecules likely interact intracellularly prior to arrival on the cell surface. Whether DISC formation and induction of the apoptosis signal occurs intracellularly remains to be proven but is suggested by these data.

MATERIALS

AND

METHODS

Cell lines. The prostate cancer cell lines PPC-1, TSU-Pr1, PC-3, LNCaP, and DU145 were maintained in RPMI 1640 media (GibcoBRL, Rockville, MD).

MOLECULAR THERAPY Vol. 2, No. 4, October 2000 Copyright  The American Society of Gene Therapy

PC-3, LNCaP and DU145 were purchased from ATCC (Rockville, MD), whereas PPC-1 and TSU-Pr1 cells were a gift from Dr. Yi Lu (University of Tennessee, Memphis, TN). The human T cell leukemia line Jurkat, a gift from Dr. Yusuf Hannun (Medical University of South Carolina, Charleston, SC) was also maintained in RPMI 1640 medium. The human chronic myelogenous leukemia cell line K-562 (ATCC) was maintained in Iscove’s modified Dulbecco’s medium (GibcoBRL). 293 cells (Microbix, Ontario, CA) were cultured in IMEM (Biofluids, Rockville, MD) containing 4 mM L-glutamine (GibcoBRL), and 293/CrmA cells in DMEM (Mediatech, Inc., Herndon, VA). All media was supplemented with 10% heat-inactivated FBS (Hyclone, Logan, UT). Cells were passaged biweekly and were incubated at 37°C with 5% CO2. The 293/CrmA packaging cell line was generated as previously described (24). Construction of plasmids and recombinant adenoviral vectors. pEGFP1-C1 (here referred to as pCMVGFP) was obtained from Clontech (Palo Alto, CA). It encodes a red-shifted gene variant of wild-type GFP that has brighter fluorescence and humanized codon usage (25). The GFPFasL fusion gene was constructed by cloning murine FasL (aa 11 to aa 279) into the C-terminus of the GFP sequence within the plasmid EGFP1-C1 and inserting the linker 5′-TCCGGCCGGACT-3′ (Arg, Gly, Arg, Thr) at the fusion site to create pGFPFasL (24). The murine FasL cDNA was generously provided by Dr. Shigekazu Nagata (Osaka Bioscience Institute). Genomic AdGFPFasLTET DNA was constructed by removing GFPFasL from pGFPFasL and inserting it into pUHD10-3 (a gift from Dr. H. Bujard), producing p10-3GFsl. The Tet-OFF fusion activation protein expression cassette was extracted from pUHD15-1 (also obtained from Dr. Bujard) and inserted into the E1 shuttle plasmid, pLAd-CMV, generating pLAd-CtTaS. GFPFasL was excised from p10-3GFsl and inserted into pLAd-CtTaS, resulting in pLAd-T.GFsL which constitutively expresses the tetracycline-controlled transactivator (tTA) and places the GFPFasL fusion under transcriptional control of the Tet response element (TRE) (26). The GFPFasL cassette was released from the shuttle plasmid pLAd-T.GFsL by digestion with SwaI and AvrII, and ligated into XbaI-digested Ad5 genomic DNA containing deletions in E1, E3, and E4 (retaining ORF6) (27). Recombinant AdGFPFasLTET virus was generated by transfecting the ligation product into 293/CrmA cells in the presence of 1 µg/ml doxycycline. Crude viral lysate was prepared when cells showed evidence of cytopathic effect. After three rounds of plaque purification, viral DNA was isolated (28) and sequenced for verification of DNA inserts. Largescale viral preparations were purified by double cesium chloride banding and the banded viral DNA was dialyzed in 10% glycerol, 10 mM Tris, 2 mM MgCl2, pH 7.5, before storing aliquots at −80°C. PCR was used to analyze stocks for RCA (29). Virus stocks were titered in triplicate using 96 well plates. Titer was determined by counting the number of GFP positive cells. Genomic AdCMVGFP DNA was constructed by cloning the cDNA for GFP into the E1 Ad5 shuttle plasmid, pLAd-CMV, generating pLAd-C.Gf. Plasmid LAd-C.Gf was digested with SwaI and AvrII releasing the GFP transgene cassette which was ligated to XbaI-digested Ad5 genomic DNA containing E1, E3, and E4 (retaining E4 ORF6) deletions (27). Recombinant AdCMVGFP virus was generated by transfecting the ligation mixture into 293 cells. Virus was prepared as described above. Flow cytometry analysis for detection of cell surface FasR. Cells were harvested at 90% confluency using a trypsin-free chelating solution, pH 7.4 (135 mM NaCl, 5 mM KCl, 20 mM Hepes buffer, and 1.5 mM EDTA). Jurkat and K-562 cells served as positive and negative controls, respectively. Cells (5 × 105) were washed once with PBS (pH 7.4, containing 0.1% sodium azide), gently resuspended in 50 µl PBS containing 2% heatinactivated FBS and 10–40 ng/µl CH-11 (Kamiya Biomedical, Seattle, WA) or IgM isotypic control (Immunotech, France), and then incubated for 1 h at 4°C. Cells were washed twice with PBS, resuspended in 50 µl PBS containing 2% heat inactivated FBS and 1.5–6 µg of goat anti-mouse PE-conjugated IgM antibody (Biomeda Corp., Foster City, CA). Both primary and secondary antibodies were titered on each cell line to ensure antibody saturation. Samples were gently vortexed and incubated in the dark for 1 h at 4°C. Cells were washed twice with PBS and analyzed by flow cytometry using a Becton–Dickinson FACSCalibur. A minimum of 10,000 events was scored for each cell line. Nonviable cells were excluded from analysis by staining with 7-amino-actinomycin D (Pharmingen, San Diego, CA).

349

ARTICLE for 24 h at 37°C with 5% CO2. Cells were then washed with PBS, trypsinized, and resuspended in 300 µl PBS. GFP expressing cells were counted using a Becton–Dickinson FACSCalibur, and marker position set using untransduced cells. A minimum of 10,000 events was scored for each cell line. MTS cytotoxicity assay. Cytotoxicity was determined with the CellTiter 96 AQueous One Solution cell proliferation assay (Promega, Madison, WI) which measures the metabolic activity of viable cells. For CH-11 antibody experiments, cells were seeded in 24-well plates, grown to 75% confluency and treated with either 500 ng/ml anti-Fas antibody clone CH-11 (Kamiya Biomedical), or 500 ng/ml normal mouse serum (Jackson Labs, West Grove, PA) for 48 h. For adenoviral transductions, 5 × 104 cells/well were infected at an m.o.i. of between 10 and 1000 (depending on the cell line) using either AdCMVGFP or AdGFPFasLTET. After 48 h of incubation, 100 µl of CellTiter 96 AQueous One Solution reagent was added per well and plates incubated for an additional 3 h. Finally, 120 µl of medium was transferred into a 96-well microtiter plate and absorbance readings were taken at 490 nm using a Vmax kinetic microplate reader (Molecular Devices, Sunnyvale, CA). Percent cytotoxicity was calculated using the formula % cytotoxicity = [1 − (OD of experimental well/OD of positive control well)] × 100. Positive controls were left untreated. Background absorbance was subtracted from each cell line using a media control. ZB-4 neutralizing antibody experiment. Cells (5 × 105) were pretreated for 1 h using 500 ng/ml of anti-Fas neutralizing antibody ZB-4 (Kamiya Biomedical) in 24-well plates. Cells were then transduced with AdCMVGFP or AdGFPFasLTET, m.o.i. 100, and assayed for cell death using the MTS assay at 24 h. TUNEL assay. Cells were stained using an APO-BRDU kit (Pharmingen). We substituted an R-phycoerytherin (R-PE)-conjugated mouse anti-BrdU monoclonal antibody (Pharmingen) for the kit’s FITC labeled antibody in order to analyze GFP fluorescence and anti-BrdU labeling simultaneously. Cells (5 × 105) were transduced in six-well dishes with either AdCMVGFP or AdGFPFasLTET. At 24 h, attached and floating cells were harvested and washed with PBS. Samples were then processed according to the manufacturer’s protocol (Apoptosis, Applied Reagents and Technologies, Instruction Manual, 2nd ed., December 1998) and sorted using a Becton–Dickinson FACSCalibur cell sorter.

FIG. 1. Cell surface Fas expression on different PCa cell lines as determined by FACS analysis. Cells were stained using the primary anti-Fas antibody CH11 followed by an IgM-phycoerytherin (PE)-conjugated secondary antibody (bold line). As a control (dotted line), an irrelevant IgM isotypic primary antibody was used followed by the IgM-PE-conjugated secondary. K-562 cells (Fas−) were used as a negative control and Jurkat cells (FasR+) were used as a positive control. Percentages represent the number of Fas expressing cells minus the control.

Immunocytochemistry and fluorescent confocal imaging. LNCaP cells were infected with AdGFPFasLTET or AdCMVGFP at an m.o.i. of 10 in 8-well Permanox-treated chamber slides. After 14 h, cells were fixed in 3.7% formalin, permeabilized with 0.2% Triton X-100 in PBS, placed in 10% normal goat serum for 20 min and then incubated with a mouse monoclonal Fas antibody (B-10, Santa Cruz, Santa Cruz, CA) at 1:50 for 1 h at room temperature. After washing with PBS, slides were incubated with a goat-anti mouse IgG antibody conjugated to rhodamine-Red-

Determining adenoviral transduction efficiency using flow cytometry. Transductions were performed in six-well plates containing 5 × 105 cells/well. Cells were seeded and transduced simultaneously in a final volume of 2 ml with AdCMVGFP at a m.o.i. of 100. The cells were incubated

TABLE 1 Fas-Mediated Cytoxicity in Prostate Cancer Cell Lines Treated with anti-Fas Antibody or AdGFPFasLTET (Expressed as % Cytoxicity  SD) Cell line

Normal mouse serum (500 ng/ml)

Anti-Fas IgM (CH-11) (500 ng/ml)

AdCMVGFP (m.o.i. 100)

AdGFPFasLTET (m.o.i. 100)

DU145

0.8  7.1

6.0  10.4

1.9  2.8

69.6  4.5

PC-3

1.3  5.8

1.3  2.2

0.9  5.0

84.8  1.1

PPC-3

2.3  0.3

29.2  2.3

2.3  6.1

98.0  7.1

LNCaP

7.5  14.2

11.6  13.7

1.6  3.2

74.9  4.3

TSU-Pr1

-3.5  2.3

-1.9  2.8

11.6  7.0

81.3  5.0

Jurkat (ctrl)

2.1  5.3

72.3  0.9

-19.5  22.5a

93.0  3.4a

K-562 ( ctrl)

—

—

-1.3  5.5a

-11.4  8.1a

Note. Percentage cytoxicity was determined using the MTS assay as described under Materials and Methods. Numbers represent two separate experiments performed in triplicate. a m.o.i. 1000.

350

MOLECULAR THERAPY Vol. 2, No. 4, October 2000 Copyright  The American Society of Gene Therapy

ARTICLE

FIG. 2. Construction of AdGFPFasLTET and AdCMVGFP. The plasmids pLAd-T.GFsL (A) and pLAd-C.Gf (B) contain the transgene cassettes GFPFasL and GFP, respectively. Both plasmids contain the left end Ad5 ITR and packaging signals (ψ). Digestion of pLAd-T.GFsL with SwaI/AvrII releases the transgene cassette, tTA-GFPFasL-TRE (C). GFP and the FasL cDNAs have been fused using a 12-base-pair linker. GFPFasL is under the transcriptional control of the tetracycline responsive element (TRE). Within C, the tetracycline-controlled transactivator (tTA) is composed of both the tetracycline repressor (TetR) and the VP16 activation domain. tTA is constitutively expressed and binds the TRE, activating GFPFasL transcription in the absence of tetracycline. Digestion of pLAd-C.Gf with SwaI/AvrII releases the GFP cassette (D). Each transgene cassette was ligated in vitro with XbaI-digested genomic Ad5 DNA (E3, E4 deleted). Recombinant adenovirus was formed by transfecting the ligation products into the respective packaging cell lines.

X (Jackson Immuno Res, West Grove, PA) at a dilution of 1:50 for 45 min. Slides were washed in PBS and coverslipped after antifade was applied. Cells were viewed at 400× using a 40× immersion lens. Images were collected on a Bio-Rad MRC 1024 laser scanning confocal microscope equipped with a krypton/argon mixed-gas laser (Zeiss, Thornwood, NY). Statistical analysis. Unless otherwise stated all assays were performed in triplicate and results were expressed as means ± SD. Statistical analysis was determined using a two-sample t test.

MOLECULAR THERAPY Vol. 2, No. 4, October 2000 Copyright  The American Society of Gene Therapy

RESULTS Human Prostate Cancer Cell Lines Express Cell Surface Fas but Are Generally Resistant to CH-11Induced Cell Death Five human prostate cancer cell lines, including TSUPR1, PPC-1, PC-3, DU145, and LNCaP, were analyzed by

351

ARTICLE

FIG. 3. PCa cells transduced in vitro with AdGFPFasLTET undergo Fas-mediated apoptosis. Each PCa cell line was transduced at an m.o.i. of 100 using both AdGFPFasLTET and AdCMVGFP, separately. Fluorescent and bright-field photographs were taken at 48 h of the same field. Fluorescent photographs indicate that cells have been transduced because they are GFP and GFPFasL expressing.

352

MOLECULAR THERAPY Vol. 2, No. 4, October 2000 Copyright  The American Society of Gene Therapy

ARTICLE TABLE 2 Adenoviral Transduction Efficiency in Human Prostate Cancer Cell Line Cell line

% of cells expressing GFP

Fold increase in geometric mean fluorescence

PPC-1

98.63

90.8

PC-3

98.43

195.0

DU145

99.14

600.0

LNCaP

94.46

42.0

TSU-Pr1

95.7

267.7

Note. "% of cells expressing GFP" reflects the percentage of GFP-positive cells in the total population. "Fold increase in geometric mean fluorescence" is a measure of GFP fluorescence intensity calculated using the following formula: geometric mean fluorescence of transduced sample/geometric mean fluorescence of untransduced sample.

FACS for levels of cell surface Fas expression using the monoclonal anti-Fas antibody CH-11. As shown in Fig. 1, all PCa cell lines examined expressed Fas on the cell surface. Various levels of cell surface Fas were detected ranging from 61.7% on DU145 cells to 93.3% on TSUPr1 cells. Fas expressing Jurkat cells were used as a positive control (99.8% positive) (30) and K-562 (Fas−) cells were used as a negative control (31). To determine whether PCa cell lines were susceptible to CH-11 induced apoptosis, cells were incubated with 500 ng/ml of CH-11 for 48 h and assayed for cell death using the MTS assay. As shown in Table 1, CH-11 killed 72.3% of Jurkat cells but of the PCa cell lines, only PPC1 cells demonstrated sensitivity to CH-11 (29.2% cell death). No correlation between level of cell surface Fas expression and sensitivity to CH-11 induced cell death was observed.

Adenovirus Transduction Efficiency in Human PCa Cells We generated a replication-deficient adenovirus construct that contains GFP driven by CMV (Figs. 2B, 2D, and 2E). This AdCMVGFP virus was used to determine transduction efficiency in the PCa and control cell lines used in this study. Cells were infected at an m.o.i. of 1 to 1000 and analyzed for GFP expression by flow cytometry at 24 h. Adenovirus-associated toxicity was determined morphologically and by staining with 7-AAD. In dying cells, membrane integrity is lost and 7-AAD can access the DNA. The results revealed that in PCa cell lines a m.o.i. of 100 was optimal for achieving maximum gene transfer while minimizing the amount of adenovirus-associated cytotoxicity. As shown in Table 2, all of the PCa cell lines tested were transduced nearly completely at m.o.i. of 100 (94.5 to 99.1%).

AdGFPFasLTET Transduction in Human PCa Cell Lines Induces Fas-Mediated Apoptosis To determine if intracellular FasL expression could induce apoptosis in PCa cells we engineered a FasL-conMOLECULAR THERAPY Vol. 2, No. 4, October 2000 Copyright  The American Society of Gene Therapy

taining adenoviral vector (AdGFPFasLTET). Murine FasL was fused to the GFP reporter (Fig. 2C), generating a fusion protein easily identified by UV illumination. The GFPFasL fusion gene was under transcriptional regulation of the Tet-Off system (26). Therefore, addition of tetracycline or one of its derivatives, i.e., doxycycline, functions to downregulate GFPFasL expression. The GFPFasL cassette was inserted into the E1 backbone region of an E1, E3, E4deleted adenovirus serotype 5 (Figs. 2A, 2C, and 2E). The ability of AdGFPFasLTET to induce Fas-mediated cell death in PCa cells was determined by transducing TSU-Pr1, PPC-1, PC-3, DU145, and LNCaP with either AdGFPFasLTET or control virus, AdCMVGFP at an m.o.i. of 100. After 48 h cell morphology was examined using fluorescent and light microscopy, and cell death was quantitated by MTS assay. Transduction with AdGFPFasLTET resulted in 70–98% cell death (Table 1), whereas control virus transduction resulted in only 0.9–11.6% cytotoxicity. GFPFasL expressing cells appeared rounded and detached from the dish by 48 h (Fig. 3). PPC-1 cells were the most sensitive, and essentially all treated cells were dead by 48 h. PC-3, DU145, TSU-Pr1, and LNCaP cells were slightly less sensitive. Although most of these cells were dead by 48 h, total cell ablation was not observed until 96 h posttreatment (data not shown). To distinguish apoptosis from nonspecific adenoviral cytotoxicity (necrosis), a TUNEL assay was performed on AdGFPFasLTET infected PPC-1 and DU145 cells. The TUNEL assay has routinely been used to detect apoptotic cells based on fragmented DNA (32). Forty-eight hours after transduction, 81.8% of PPC-1 and 35.2% of DU145 cells stained positive for TUNEL (Fig. 4). These data indicate that AdGFPFasLTET transduced cells undergo apoptosis and not necrosis. Next we determined that AdGFPFasLTET-induced cytotoxicity was indeed mediated by Fas. Fas+ (Jurkat) and Fas− (K-562) cell lines were transduced using either AdGFPFasLTET or AdCMVGFP. These control cell lines were difficult to infect necessitating the use of an m.o.i. of 1000 to realize detectable gene transduction. As expected, treatment of Jurkat and K-562 cells with AdCMVGFP resulted in no adenovirus-associated cytotoxicity as determined by both MTS assay (Table 1) and 7-AAD dye exclusion. FACS analysis detected 75% of AdCMVGFP infected Jurkat cells expressing GFP and only 3.5% of these cells stained positive for 7-AAD at 48 h. In K-562 cells, 81% of the cells were GFP expressing and 10.5% stained positive for 7-AAD (48 h). In contrast, AdGFPFasLTET treatment resulted in cytotoxicity of Fas+ Jurkat cells, but not Fas− K-562 cells. FACS analysis detected 48% of Jurkat cells staining positive for 7-AAD whereas only 9.8% of K-562 cells stained positive. In addition, the MTS assay detected 93% cell death (Table 1) in Jurkat cells and no cell death in K-562 cells. If AdGFPFasLTET-induced cell death is specific for GFPFasL, it should be possible to modulate cytotoxicity with tetracycline. Therefore, DU145 cell were infected with AdGFPFasLTET at an m.o.i. of 100 in the presence or absence of 10 µM tetracycline and cell death was meas-

353

ARTICLE TABLE 3 Effect of AdCMVGFP Transduction on CH-11-Induced Apoptosis in PCa Cells Cell line

Normal mouse serum (500 ng/ml)

DU145

CH-11 (500 ng/ml)

AdCMVGFP (m.o.i. 100)

Cell death (MTS assay) %





3.3  4.4



6.0  2.4

  

PPC-1 

8.1  8.1  



12.3  7.4 4.4  4.5 31.1  0.6

AdCMVGFP Treatment in DU145 and PPC-1 Cells Does Not Increase CH-11 Sensitivity Next we wanted to determine if a recombinant adenovirus gene/protein was responsible for conferring Fassensitivity to CH-11 resistant PCa cells. DU-145 (CH-11 resistant) and PPC-1 cells (partially CH-11 resistant) were pretreated with AdCMVGFP (m.o.i. 100) on day 1, then challenged on day 2 with CH-11 (500 ng/ml). Cells were confirmed to be GFP expressing using fluorescent microscopy. On day 3 MTS assay revealed a decrease in CH-11-associated cytotoxicity in both DU-145 and PPC1 cells (Table 3), suggesting that recombinant adenovirus does not confer Fas sensitivity.

Anti-Fas Neutralizing Antibody ZB-4 Does Not Block AdGFPFasLTET-Associated Cytotoxicity The anti-Fas neutralizing antibody ZB-4 has previously been used to block the cytotoxic effects of Fas ago-

AdGFPFasL(TET)

AdCMVGFP

FIG. 4. PPC-1 and DU145 cells transduced with AdGFPFasLTET undergo apoptosis as determined by TUNEL assay. PPC-1 cells were transduced at an m.o.i. of 10 and DU145 cells at an m.o.i. of 100 with AdGFPFasLTET and AdCMVGFP, respectively. Cells were stained at 24 h using a PE conjugated anti-BrdU antibody (for control a PE conjugated isotypic antibody was used), and then examined by FACS analysis. Results are expressed as percent antiBrdU positive over isotypic control. (A) In PPC-1 cells, 83% of the cells stained positive for BrdU while 61% of the cells had detectable GFPFasL. PPC-1 cells are very sensitive to AdGFPFasLTET treatment therefore some cells die before GFPFasL expression can be measured by fluorescence. In AdCMVGFP treatment, 93% of the cells were expressing GFP and only 3% of the cells stained positive for BrdU. (B) In DU145 cells, 35.2% of the cells stained positive for BrdU, while 98% of the cells were expressing GFPFasL. In the AdCMVGFP treatment, 91.8% of the cells were expressing GFP and only 0.3% of the cells stained positive for BrdU.

ured at 48 h using MTS assay. Tetracycline treatment reduced AdGFPFasLTET-associated cytotoxicity by 54.4%, from 77.8 ± 0.35 to 35.5 ± 4.5% but was not able to completely block GFPFasL-induced cell death (data not shown and see Ref. 23).

354

FIG. 5. Anti-Fas blocking antibody ZB-4 does not block AdGFPFasLTETassociated cytotoxicity in LNCaP and PPC-1 cells. PPC-1 cells were pretreated for 1 h using 500 ng/ml ZB-4 followed by the anti-Fas agonistic antibody CH11 (500 ng/ml). ZB-4 treatment reduced CH-11 cytotoxicity from 39.7 to 3.1% (P < 0.005). Next we examined if ZB-4 was able to block the cytotoxic effects of AdGFPFasLTET. LNCaP and PPC-1 cells were pretreated for 1 h with ZB-4 (500 ng/ml), and then challenged with AdGFPFasLTET or AdCMVGFP (m.o.i. 100). After 24 h, MTS assay indicated minimal blockage in PPC-1 cells and no blockage in LNCaP cells. MOLECULAR THERAPY Vol. 2, No. 4, October 2000 Copyright  The American Society of Gene Therapy

GFP/Fas

GFPFasL/Fas

Fas

Fas

GFP

GFPFasL

ARTICLE

AdGFPFasL(TET)

AdCMVGFP

FIG. 6. Intracellular protein localization of GFPFasL, GFP, and Fas in LNCaP cells transduced with AdGFPFasLTET and AdCMVGFP, respectively, using an m.o.i. of 10. Cells were immunostained with a Fas antibody as described under Materials and Methods 14 h after viral treatment. Images shown are a stack of 20 images captured using laser scanning confocal microscopy. In AdCMVGFP treated cells (D), diffuse GFP expression is seen both in the cell’s cytoplasm and in the nucleus. In contrast, GFPFasL expression (A) has been excluded from the nucleus (arrow indicates nuclear ghost, faint staining in this area is from GFPFasL expression on the plasma membrane), and is seen in a punctated pattern within the cytoplasm. Fas expression (B and E) is seen both on the membrane and punctated within the cytoplasm (arrow in B indicates a nontransduced cell). GFPFasL and Fas (C) colocalize to the same intracellular compartment [arrow indicates an area of yellow color where GFPFasL (green) and Fas (red) are coexpressed]. This colocalization area is likely the ER, Golgi, and/or secretory vesicles. Images were captured at 400× using immersion optics. MOLECULAR THERAPY Vol. 2, No. 4, October 2000 Copyright  The American Society of Gene Therapy

355

ARTICLE nistic antibodies (33), and FasL expressing CTLs (34). Pretreatment using ZB-4 (500 ng/ml) followed by CH-11 (500 ng/ml) resulted in almost complete blockage of CH11-induced cell death (Fig. 5). Next we determined if ZB4 was able to block AdGFPFasLTET-induced cytotoxicity in PCa cells. LNCaP and PPC-1 cells were pretreated for 1 h using ZB-4 and then challenged with AdGFPFasLTET or AdCMVGFP (m.o.i. 100). After 24 h, the MTS assay indicated minimal blockage (12.5%) in PPC-1 cells and no blockage in LNCaP cells. This data suggests that GFPFasL and Fas interact intracellularly prior to arrival at the cell’s surface and leads to initiation of the death signal before ZB-4 has an opportunity to block signaling.

Cellular Localization of Fas, GFP, and GFPFasL To determine if Fas and GFPFasL colocalize to the same intracellular compartment LNCaP cells were transduced with either AdGFPFasLTET or AdCMVGFP and then immunostained with a Fas antibody as described under Materials and Methods at 14 h posttransduction. Confocal microscopy was used to analyze the expression patterns of Fas, GFP, and GFPFasL. In most cells, GFP was evenly distributed throughout the cytoplasm and nucleus (Fig. 6D) (some cells had more GFP expression in the nucleus, data not shown). In contrast, GFPFasL was excluded from the nucleus and seen in the cytoplasm and plasma membrane (Fig. 6A). Punctated GFPFasL and Fas expression was seen in both the plasma membrane and within what is likely the Golgi and/or endoplasmic reticulum (Figs. 6B and 6E). Many cells showed colocalization of GFPFasL and Fas within the same intracellular compartment (Fig. 6C).

DISCUSSION Efforts to develop novel therapies for the treatment of advanced PCa are underway in many laboratories. One promising therapeutic approach is to induce apoptosis in cancer cells. The cell surface receptor Fas is mostly known for its role in inducing apoptosis in the immune system but it may also play a role in the prostate (19). Although the Fas/FasL pathway is functional in PCa cells, most Fas expressing cell lines are resistant to induction of apoptosis by agonistic Fas antibodies, such as CH-11 (14) or IPO-4 (35). However, our results suggest that the apoptotic potential of these cells may have been underestimated. In this study, we have generated an adenoviral construct, AdGFPFasLTET that expresses a GFPFasL fusion protein under the control of a tetracycline-regulatable promoter. The AdGFPFasLTET construct consistently killed 70–98% of cells 48 h after transduction. Cell death was apoptotic as demonstrated by TUNEL and we demonstrated that in order for the AdGFPFasLTET to be effective, cells have to express Fas. When DU145 and PPC-1 cells were infected with a AdCMVGFP control virus and then challenged with CH-

356

11, cell death was not augmented, indicating that an Ad5 gene or protein was unlikely to be responsible for sensitizing these cells to Fas-mediated apoptosis. Taken together, these results indicate that AdGFPFasLTETinduced apoptosis was specific for the transgene. Finally, we have previously published (24) that attenuation of AdGFPFasLTET-associated cytotoxicity was achieved by addition of tetracycline, demonstrating that we have a controllable system. Our results suggest that AdGFPFasLTET is more toxic than an adenovirus with a CMV-driven murine FasL adenovirus (Ad5d/327CMV-mFasL) (36). Although both adenovirus constructs induced similar toxicity in PPC-1 cells, our construct caused a more dramatic response in DU145, LNCaP and TSU-Pr1 cells. These differences may be attributed to the different assays used to detect cytotoxicity. Alternatively, the difference may lie with the GFPFasL fusion protein, which is predicted to be more stable (under investigation) than FasL alone. We also constructed a virus expressing only mouse FasL and found this virus to be less effective than the GFPFasL construct (unpublished). The mechanism by which Ad5d/327CMV-mFasL or AdGFPFasLTET overcomes resistance to apoptosis induced by agonistic Fas-antibodies such as CH-11 or IPO-4 is unknown. One theory suggests that CH-11 and endogenous ligand may activate different pathways. Thilenius et al. found that CH-11 and natural FasL stimulated different signaling pathways in EL4 cells (36). Alternatively, it has been hypothesized that an antiapoptotic protein is responsible for blocking the Fas pathway. In those studies, the CH-11 antibody failed to induce apoptosis in PC-3 and DU145 cells but pretreatment of these cells with subtoxic levels of certain chemotherapeutic drugs (CDDP, VP-16, or ADR), followed by CH-11 treatment, did result in Fas-mediated apoptosis (34). This suggests that the apoptotic machinery is still in place and that treatment with chemotherapeutic drugs removes a blockage in the Fas pathway. If this hypothesis is correct, internal GFPFasL expression may circumvent this blockage and induce Fas-mediated apoptosis in the absence of chemotherapeutic drugs. Perhaps future studies will reveal the underlying mechanism for overcoming Fas resistance. Surprisingly, we were unable to prevent AdGFPFasLTET-induced apoptosis with the neutralizing antibody ZB4. These results suggest that GFPFasL may be complexing with Fas intracellularly circumventing the inhibitory action of externally applied ZB4. An alternative is that ZB4 is unable to disrupt DISC formation because DISC assembly is complete upon arrival at the cell surface. To address this question we examined the location of GFP, GFPFasL, and Fas by confocal microscopy. GFP expression is seen uniformly distributed throughout the cell (Fig. 6D) however, when GFP is fused to murine FasL a very different expression pattern is observed (Fig. 6A). GFPFasL expression is seen on the plasma membrane and punctated throughout the MOLECULAR THERAPY Vol. 2, No. 4, October 2000 Copyright  The American Society of Gene Therapy

ARTICLE cytoplasm (Fig. 6C). Using confocal microscopy we show that Fas and GFPFasL colocalize to the same intracellular compartment. We suspect colocalization to occur in the Golgi complex based on three previous reports localizing Fas (23), TRAIL (37), and Fas Ligand (43) to the Golgi complex. Currently we are investigating the identity of this intracellular compartment in our system. It should be noted that some GFPFasL protein may also be targeted to the proteosome and/or aggresome (38, 39). From these data we postulate that intracellular overexpression of GFPFasL leads to a ligation event within the ER, Golgi, and/or secretory vesicles which initiates DISC formation much earlier than previously expected. Internal activation of a plasma membrane receptor has previously been demonstrated using a v-sis encoded mitogen that binds the PDGF receptor intracellularly, although activated receptor still had to localize to the cell surface for signaling to occur (40). In the context of gene therapy, using FasL as a molecular therapeutic approach has several advantages over current strategies, perhaps most importantly is that it takes advantage of the bystander effect. Previously, it has been demonstrated that FasL expressing cells induce apoptosis in Fas expressing prostate (21, 22) and nonprostate (41) target cells. It has further been shown that apoptotic cells release vesicles that continue to present FasL and can initiate apoptosis (42). These observations have relevance to our system since we observed that the number of GFPFasL expressing cells is usually less than the number of cells undergoing apoptosis (24) (the details of this bystander effect are currently under investigation). In conclusion, our results indicate that intracellular GFPFasL expression is superior to the Fas agonist CH-11 at inducing Fas-mediated apoptosis in the PCa cell lines examined. We have demonstrated that AdGFPFasLTET is capable of inducing between 70 and 98% cell death in all PCa cell lines studied to date. In the future, we will evaluate the use of AdGFPFasLTET as a potential gene therapy for human PCa. We anticipate that coupling the cytotoxic effects of GFPFasL with a prostate-specific promoter will lead to a safe bystander gene therapy approach for prostate cancer. ACKNOWLEDGMENTS We thank Drs. Margaret Kelly, David A. Schwartz, and Dennis Watson for their help in preparing the manuscript. We thank Drs. Bobby Thompson and Steve Kubalak for capturing the confocal images. Finally, we thank Candace Enockson for acquiring all the FACS data with the use of the MUSC Flow Cytometry Facility equipment that is funded by the Hollings Cancer Center. This study was supported by NIH R01 CA69596.

REFERENCES 1

Landis, S. H., Murray, T., Bolden, S., and Wingo, P. A. (1999). Cancer statistics [see comments]. CA Cancer J. Clin. 49: 8–31. 2 Konety, B. R., and Getzenberg, R. H. (1997). Novel therapies for advanced prostate cancer. Semin. Urol. Oncol. 15: 33–42. 3 Rodriguez, R., Schuur, E. R., Lim, H. Y., Henderson, G. A., Simons, J. W., and

MOLECULAR THERAPY Vol. 2, No. 4, October 2000 Copyright  The American Society of Gene Therapy

Henderson, D. R. (1997). Prostate attenuated replication competent adenovirus (ARCA) CN706: A selective cytotoxic for prostate-specific antigen-positive prostate cancer cells. Cancer Res. 57: 2559–2563. 4 Hrouda, D., Perry, M., and Dalgleish, A. G. (1999). Gene therapy for prostate cancer. Semin. Oncol. 26: 455–471. 5 Takahashi, T., Tanaka, M., Inazawa, J., Abe, T., Suda, T., and Nagata, S. (1994). Human Fas ligand: gene structure, chromosomal location and species specificity. Int. Immunol. 6: 1567–1574. 6 Suda, T., and Nagata, S. (1994). Purification and characterization of the Fas-ligand that induces apoptosis. J. Exp. Med. 179: 873–879. 7 Boldin, M. P., Varfolomeev, E. E., Pancer, Z., Mett, I. L., Camonis, J. H., and Wallach, D. (1995). A novel protein that interacts with the death domain of Fas/APO1 contains a sequence motif related to the death domain. J. Biol. Chem. 270: 7795–7798. 8 Chinnaiyan, A. M., O’Rourke, K., Tewari, M., and Dixit, V. M. (1995). FADD, a novel death domain-containing protein, interacts with the death domain of Fas and initiates apoptosis. Cell 81: 505–512. 9 Boldin, M. P., Goncharov, T. M., Goltsev, Y. V., and Wallach, D. (1996). Involvement of MACH, a novel MORT1/FADD-interacting protease, in Fas/APO-1- and TNF receptorinduced cell death. Cell 85: 803–815. 10 Muzio, M., et al. (1996). FLICE, a novel FADD-homologous ICE/CED-3-like protease, is recruited to the CD95 (Fas/APO-1) death-inducing signaling complex. Cell 85: 817–827. 11 Sasaki, Y., Ahmed, H., Takeuchi, T., Moriyama, N., and Kawabe, K. (1998). Immunohistochemical study of Fas, Fas ligand and interleukin-1 beta converting enzyme expression in human prostatic cancer. Br. J. Urol. 81: 852–855. 12 Xerri, L., Devilard, E., Hassoun, J., Mawas, C., and Birg, F. (1997). Fas ligand is not only expressed in immune privileged human organs but is also coexpressed with Fas in various epithelial tissues. Mol. Pathol. 50: 87–91. 13 Liu, Q. Y., Rubin, M. A., Omene, C., Lederman, S., and Stein, C. A. (1998). Fas ligand is constitutively secreted by prostate cancer cells in vitro. Clin. Cancer Res. 4: 1803–1811. 14 Hedlund, T. E., Duke, R. C., Schleicher, M. S., and Miller, G. J. (1998). Fas-mediated apoptosis in seven human prostate cancer cell lines: Correlation with tumor stage. Prostate 36: 92–101. 15 Griffith, T. S., Brunner, T., Fletcher, S. M., Green, D. R., and Ferguson, T. A. (1995). Fas ligand-induced apoptosis as a mechanism of immune privilege [see comments]. Science 270: 1189–1192. 16 Bellgrau, D., Gold, D., Selawry, H., Moore, J., Franzusoff, A., and Duke, R. C. (1995). A role for CD95 ligand in preventing graft rejection [see comments] [published erratum appears in Nature 394(6689):133, 1998]. Nature 377: 630–632. 17 Stuart, P. M., Griffith, T. S., Usui, N., Pepose, J., Yu, X., and Ferguson, T. A. (1997). CD95 ligand (FasL)-induced apoptosis is necessary for corneal allograft survival. J. Clin. Invest. 99: 396–402. 18 French, L. E., et al. (1996). Fas and Fas ligand in embryos and adult mice: Ligand expression in several immune-privileged tissues and coexpression in adult tissues characterized by apoptotic cell turnover. J. Cell Biol. 133: 335–343. 19 Suzuki, A., Matsuzawa, A., and Iguchi, T. (1996). Down regulation of Bcl-2 is the first step on Fas-mediated apoptosis of male reproductive tract. Oncogene 13: 31–37. 20 Watanabe-Fukunaga, R., Brannan, C. I., Copeland, N. G., Jenkins, N. A., and Nagata, S. (1992). Lymphoproliferation disorder in mice explained by defects in Fas antigen that mediates apoptosis. Nature 356: 314–317. 21 Frost, P., Ng, C. P., Belldegrun, A., and Bonavida, B. (1997). Immunosensitization of prostate carcinoma cell lines for lymphocytes (CTL, TIL, LAK)-mediated apoptosis via the Fas–Fas-ligand pathway of cytotoxicity. Cell. Immunol. 180: 70–83. 22 Hedlund, T. E., et al. (1999). Adenovirus-mediated expression of Fas ligand induces apoptosis of human prostate cancer cells. Cell Death Differ. 6: 175–182. 23 Bennett, M., Macdonald, K., Chan, S. W., Luzio, J. P., Simari, R., and Weissberg, P. (1998). Cell surface trafficking of Fas: A rapid mechanism of p53-mediated apoptosis [see comments]. Science 282: 290–293. 24 Rubinchik, S., Ding, R., Qiu, A., Zhang, F., and Dong, J. (2000). Adenoviral vector which delivers FasL–GFP fusion protein regulated by the tet-inducible expression system. Gene Ther. 7: 875–885. 25 Zhang, G., Gurtu, V., and Kain, S. R. (1996). An enhanced green fluorescent protein allows sensitive detection of gene transfer in mammalian cells. Biochem. Biophys. Res. Commun. 227: 707–711. 26 Gossen, M., and Bujard, H. (1992). Tight control of gene expression in mammalian cells by tetracycline-responsive promoters. Proc. Natl. Acad. Sci. USA 89: 5547–5551. 27 Huang, M. M., and Hearing, P. (1989). The adenovirus early region 4 open reading frame 6/7 protein regulates the DNA binding activity of the cellular transcription factor, E2F, through a direct complex. Genes Dev. 3: 1699–1710. 28 Deryckere, F., and Burgert, H. G. (1997). Rapid method for preparing adenovirus DNA. Biotechniques 22: 868–870. 29 Zhang, W. W., Koch, P. E., and Roth, J. A. (1995). Detection of wild-type contamination in a recombinant adenoviral preparation by PCR. Biotechniques 18: 444–447. 30 McGahon, A. J., Nishioka, W. K., Martin, S. J., Mahboubi, A., Cotter, T. G., and Green, D. R. (1995). Regulation of the Fas apoptotic cell death pathway by Abl. J. Biol. Chem. 270: 22625–22631. 31 Takahashi, S., et al. (1993). Establishment of apoptosis-inducing monoclonal antibody 2D1 and 2D1-resistant variants of human T cell lines. Eur. J. Immunol. 23: 1935–1941. 32 Darzynkiewicz, Z., Juan, G., Li, X., Gorczyca, W., Murakami, T., and Traganos, F. (1997). Cytometry in cell necrobiology: Analysis of apoptosis and accidental cell death

357

ARTICLE (necrosis). Cytometry 27: 1–20. 33 Yonehara, S., et al. (1994). Involvement of apoptosis antigen Fas in clonal deletion of human thymocytes. Int. Immunol. 6: 1849–1856. 34 Uslu, R., et al. (1997). Chemosensitization of human prostate carcinoma cell lines to anti-fas-mediated cytotoxicity and apoptosis. Clin. Cancer Res. 3: 963–972. 35 Rokhlin, O. W., et al. (1997). Fas-mediated apoptosis in human prostatic carcinoma cell lines. Cancer Res. 57: 1758–1768. 36 Thilenius, A. R., Braun, K., and Russell, J. H. (1997). Agonist antibody and Fas ligand mediate different sensitivity to death in the signaling pathways of Fas and cytoplasmic mutants. Eur. J. Immunol. 27: 1108–1114. 37 Zhang, X. D., Franco, A. V., Nguyen, T., Gray, C. P., and Hersey, P. (2000). Differential localization and regulation of death and decoy receptors for TNF-related apoptosis-inducing ligand (TRAIL) in human melanoma cells. J. Immunol. 164: 3961–3970. 38 Garcia-Mata, R., Bebok, Z., Sorscher, E. J., and Sztul, E. S. (1999). Characterization and dynamics of aggresome formation by a cytosolic GFP-chimera. J. Cell Biol. 146:

358

1239–1254. 39 Johnston, J. A., Ward, C. L., and Kopito, R. R. (1998). Aggresomes: A cellular response to misfolded proteins. J. Cell Biol. 143: 1883–1898. 40 Fleming, T. P., Matsui, T., Molloy, C. J., Robbins, K. C., and Aaronson, S. A. (1989). Autocrine mechanism for v-sis transformation requires cell surface localization of internally activated growth factor receptors. Proc. Natl. Acad. Sci. USA 86: 8063–8067. 41 Muruve, D. A., Nicolson, A. G., Manfro, R. C., Strom, T. B., Sukhatme, V. P., and Libermann, T. A. (1997). Adenovirus-mediated expression of Fas ligand induces hepatic apoptosis after systemic administration and apoptosis of ex vivo-infected pancreatic islet allografts and isografts. Hum. Gene Ther. 8: 955–963. 42 Martinez-Lorenzo, M. J., et al. (1999). Activated human T cells release bioactive Fas ligand and APO2 ligand in microvesicles. J. Immunol. 163: 1274–1281. 43 Yamashita, H., Otsuki, Y., Ito, Y., Matsumoto, K., Ueki, K., and Ueki, M. (1999). Fas ligand, Fas antigen and Bcl-2 expression in human endometrium during the menstrual cycle. Mol. Hum. Reprod. 5(4): 358–364.

MOLECULAR THERAPY Vol. 2, No. 4, October 2000 Copyright  The American Society of Gene Therapy