As solid tumors expand, they can rapidly outgrow the carrying capacity of the local vasculature. Tumors are, therefore, often filled with areas, characterized by lowered oxygen content, or hypoxia, and limited nutrient supply.1 These effects have many therapeutic ramifications.2 The balance between opposing response pathways elicited in the nutrient-deprived malignant cells determines the fate of the stressed malignant tumors. Characterization of cellular regulators whose function is modified under imposed stress conditions should, therefore, further our understanding of processes that govern the progression of solid tumors.
TATA element modulatory factor (TMF) is a Golgi-associated protein3 whose downregulation leads to fragmentation of the Golgi complex in rat NRK cells.4 TMF is also involved in Rab6-dependent retrograde membrane traffic.5 However, other characteristics of TMF predict its involvement in additional cellular processes. TMF was originally identified as a TATA element modulatory factor6 that binds to the androgen receptor and in PC3 prostate carcinoma cells potentiates the transcription-inducing activity of the androgen receptor in a ligand-dependent manner.7 Finally, TMF was shown to become dispersed in the cytoplasm and to mediate the delivery of Stat3 and TRNP to proteasomal degradation under conditions of growth factor deprivation.8, 9 The TMF-directed ubiquitination of Stat3 is mediated by a “BC box” element, which resides in the N-terminal portion of the protein and enables its association with an E3 ubiquitin ligase complex.8
The TMF-mediated downregulation of a key pro-oncogenic transcription factor like Stat3 could explain the reduced level of TMF/ARA160 in malignant brain tumors.8 However, TMF/ARA160 might direct the degradation of additional key transcription factors under imposed stress conditions, thereby affecting the progression of solid tumors. Furthermore, the ability of TMF/ARA160 to suppress the growth of solid tumors in vivo has not been documented. To examine whether TMF attenuates the progression of tumors and to explore novel TMF-regulated factors, TMF/ARA160 was ectopically expressed in the human PC3 prostate carcinoma cells. Although these cells efficiently grow as xenografts in athymic mice, they do not express Stat3.10 The absence of Stat3 activity turned the PC3 system suitable for identifying additional regulatory pathways, which are affected by TMF during tumor progression. Characterization of the TMF system has ramifications on the possible manipulation and management of solid tumor progression.
Material and methods
Construction of plasmids
Construction of the pcDNA3-HA-human TMF, pcDNA3-HA-TMF-ΔBC (which lacks the BC box element) was performed as described before.8, 11 The pCAGGS-MYC-ubiquitin plasmid was kindly provided by Aaron Ciechanover (Technion, Israel). The pECE-Flag-mp65-6×His was constructed in 2 consecutive steps. First, a fragment of the mouse p65 cDNA extending from nucleotide 346 to nucleotide 1995 was digested from pRSET-6×His-mp65 vector by HindIII and XhoI, purified by agarose gel and then amplified by polymerase chain reaction (PCR) with a forward primer 5′GCTCGAAGCTTGACGATCTGTTTCCCCTCATCT3′ bearing a HindIII site at its 5′ end, without the ATG codon, and a backward primer 5′CGCGGGAATTCTTAATGGTGATGGTGA TGATGGGAGCTGATCTGACTCAA3′ bearing a 6×His tag and EcoRI site at its 5′ end. The amplification product was then cut with HindIII and EcoRI and ligated into the pECE-Flag vector.
Cell cultures and treatments
Prostate carcinoma cells (PC3) were grown in RPMI (Gibco, Invitrogen, Paisley, UK) supplemented with 10% fetal calf serum (FCS) (Biological Industries, Beit Haemek, Israel). The PC-3 was initiated from a bone metastasis of a grade IV prostatic adenocarcinoma from a 62-year-old white man.
For amino acid starvation, cells were washed twice with Hank's buffered salt solution (HBSS) (Biological Industries, Beit Haemek, Israel) and maintained in HBSS for 18–20 hr. Hypoxic conditions were induced by maintaining cells in anaerobic culture jars using carbon dioxide generating envelopes (AnaeroGen™ AN0025A OXOID, Gamidordiagnostics, Basingstoke, UK), which reduced the oxygen level in the jar to below 1% within 30 min. Cells were then incubated for 20 hr under these hypoxic conditions before harvesting. For inhibition of proteasome activity, cells were grown in the presence of 35 μM of the proteasome inhibitor MG-132 (Calbiochem, San Diego, CA) for 5 hr before harvesting.
Transfection of cells
For stable transfection, PC3 cells were grown to ∼ 80% confluence in 10-cm dishes. The cells were transfected using the LipofectAMINE reagent (Invitrogen, Paisley, UK). Each transfection was performed by mixing 10 μg DNA, pcDNA3 empty vector, or pcDNA3-HA-TMF or pcDNA3-HA-TMF-ΔBC (as described previously) with 25 μL of LipofectAMINE and 1 mL serum-free Opti-MEM (Gibco, Invitrogen, Paisley, UK) medium as described before.9 After 24 hr, cells were trypsinized (Gibco, Invitrogen, Paisley, UK) and replated at 2 different dilutions (20×, 40×), in quadruplicates in 10-cm dishes. After 24 hr, selection medium consisted of standard medium supplemented with 800 mg/mL geneticin (G418, Calbiochem, San Diego, CA) was added to the cells. Selection was performed for 4 weeks. Cells grew as colonies on the dishes used for transfection, and several colonies were selected and isolated. Of 24 colonies isolated and tested for expression by Western blot, 12 could be maintained as cell lines. These cell lines were propagated until at least 10 aliquots of each cell line had been stored in liquid nitrogen. After selection, cells were maintained in standard medium supplemented with 800 mg/L geneticin (G418, Calbiochem, San Diego, CA).
For in vivo ubiquitination assay, the clones (3 × 105) were plated in 10-cm dishes, 24 hr before transfection. The cells were transfected using the Trans-IT Prostate reagent (Mirus, Madison, USA). Each transfection was performed by mixing 35 μL of Trans-IT reagent with 1 mL serum-free Opti-MEM (Gibco, Invitrogen, Paisley, UK) medium for 25 min, then 10 μg DNA was added and incubated for 25 min, at last 35 μL of boost reagent was add for 25 min. The mixture was incubated at room temperature. DNA suspensions were overlaid onto the cells. Cells were incubated for 24 hr and then subjected to starvation for 18 hr.
Flow cytometry analysis
Flow cytometry analysis was performed essentially as described before.9
Quantification of viable cells (MTT)
Quantification of viable cells was performed essentially as described before.9 Transected clones were plated at 1–2 × 103 cells per well in 96-well plates. The detection was performed 24, 48, 72 and 96 hr after plating.
Establishment of xenografts
All animal experiments were performed according to the guidelines of the Institutional Animal Care and Use Committee. Nude (BALB/c nu/nu) male mice were obtained from Harlen Laboratories (Harlen Co., Rehovot, Israel). Mice were housed in cages (4 mice/cage) under conditions of constant photoperiod (12 L:12 D) with free access to food and water. PC3 cell lines stably expressing TMF or an empty vector as control were grown in RPMI medium supplemented with 10% heat-inactivated FCS (Biological Industries, Beit Haemek, Israel). The subcutaneous implantation was performed as described.1 Briefly, 1 × 106 cells in 0.1 mL Matrigel (BD Biosciences, Bedford, MA) was inoculated subcutaneously using a 25-gauge needle into the dorsal side of 5–6-week-old nude male mice. Tumor size was measured 8 days after cell implantation, twice a week, by caliper using the formula: length × width × depth × 0.5236,12 during a period of 25 days. At the end of the experiment, the mice were sacrificed and tumors were removed and fixed.
Western blot analysis
Preparation of cell lysates and Western blot analysis was performed essentially as described before.9 Insoluble proteins were recovered in Laemmli sample buffer and protein concentration was determined as follows: Whatman 3MM paper was cut into comb shape, with 0.5–10 cm comb teeth, 0.1-cm spaces between each comb tooth. Bovine serum albumin (BSA) standards and protein samples in Laemmli buffer were loaded on the comb teeth, dried and then immersed in Coomassie brilliant blue G 250 (Merck, Darmstadt, Germany) stain solution (2.5 g Coomassie in 100 mL acetic acid, 400 mL methanol and 500 mL water) for 30 min with gentle agitation. Distaining was performed with distain solution (70 mL acetic acid, 200 mL methanol, 730 mL water) followed by drying. Each tooth was introduced into a well of 24-well plate. Color was eluted in 0.5 mL 3% SDS solution over night in RT and intensity determined by absorbance in 595 nm.
Electro blotted proteins were detected using monoclonal anti-HA antibodies (MMS-101R, Covance, CA), monoclonal antiactin antibodies (Sigma, Steinheim, Germany), polyclonal nuclear factor kappaB (NFκB) p65 (sc-109, Santa Cruz, CA), polyclonal NFκB p50 (c19, Delta laboratories, Somersby, Australia), polyclonal anti-His-probe (Santa Cruz, CA), polyclonal anti-E2F4 (sc-866, Santa Cruz, CA), polyclonal anti-AP2α (sc-184, Santa Cruz, CA), polyclonal anti-TMF antibodies8 and monoclonal anti-FLAG (Sigma, Steinheim, Germany). Immunoabsorbed antibodies were visualized using the chemiluminescence reaction (PIERCE, Rockford, USA). The blots were scanned using a Canon Scan N670U optical scanner, and the optical density of each band was determined using the ImageJ software application. For quantification, the control of each experiment was set to 100% after normalization to actin level.
Immunocytochemistry was performed essentially as described before.9 PC3 cells were exposed to 1:100 polyclonal rabbit anti-TMF or 1:100 polyclonal mouse anti-TMF, anti-HA (MMS-101R, Covance, CA) and 1:400 TGN46 (Abcam, Cambridge, UK) antibodies. Rabbit polyclonal or mouse monoclonal antibodies were visualized with Alexa Fluor 488 goat anti-rabbit/mouse secondary antibodies (Molecular Probes, Invitrogen, Paisley, UK). Bound fluorophors were visualized using a Bio-Rad (Hercules, CA) MRC 1024 upright confocal microscope with a krypton-argon ion laser. Confocal images analysis was performed using Bio-Rad software, and figures were compiled using the Laser Sharp 3.0 software package.
Immunohistochemical analysis (frozen sections)
For preparation of frozen sections, xenografts were excised, fixed in 4% paraformaldehyde overnight, rehydrated overnight in 20% sucrose, and fixed in O.C.T (TISSUE TK, Elkhart, USA). Sections of 5 μm thickness on slides were processed for immunohistochemistry using the following primary antibodies and dilutions: 1:100 proliferating cell nuclear antigen (PCNA) (AB-1 clone: PC10 NeoMarkers, Fremont, USA), 1:100 polyclonal anticleaved caspase 3 (Cell signaling, MA), rat anti mouse CD-31 (BD Pharmingen, Bedford, MA) and 1:50 anti-HA (Covance, CA). Sections were incubated in 3% H2O2 for 10 min at room temperature to quench endogenous peroxidases and then processed for antigen retrieval by incubating in 0.1% sodium citrate buffer (pH 6) for 10 min in a subboiling water bath in a microwave oven. The sections were washed 3 times. The sections were then incubated in blocking solution (5% FCS, 1% BSA, 0.5% Triton in 1× phosphate-buffered saline) for 60 min at room temperature, and then incubated overnight at 4°C with primary antibody diluted in blocking solution. The next day, sections were incubated with the biotinylated secondary antibody for 1 hr at room temperature, washed 3 times, incubated 30 min with streptavidin-HRP (Zymed, CA), rewashed, and developed with AEC staining kit (Sigma, Steinheim, Germany) for 5–20 min until staining appeared. The slides were counterstained with hematoxylin, dehydrated, and mounted with coverslips. All washes were for 5 min in 1× PBST wash buffer (1× phosphate-buffered saline with 0.2% Tween-20). The fluorescent staining sections were viewed with Olympus CX-40 (Olympus Optical, Tokyo, Japan) and photographed using a gray scale digital camera CFW-1312M (Scion, Fredrickburg, VA). Images were enhanced using PhotoShop software. The enzymatic histological staining sections were viewed with Olympus CX-40 (Olympus Optical, Tokyo, Japan) and photographed using a color digital camera CFW-1612C (Scion, Fredrickburg, VA).
For fluorescent staining with anti-HA and anti-PCNA, the secondary antibody was anti-mouse Alexa 488 (Molecular Probes, Invitrogen, Paisley, UK) and the nuclei were stained with 0.05 mg/mL propidium iodide. Bound fluorophors were detected with a Bio-Rad (Hercules, CA) MRC 1024 upright confocal microscope with a krypton-argon ion laser. Confocal microscope image analysis was performed using Bio-Rad software, and figures were compiled using the Laser Sharp 3.0 software package.
Quantification of CD-31 staining
All sections were evaluated under a standard light microscope Olympus CX-40 (Olympus Optical, Tokyo, Japan) and photographed using a monochrome digital camera CFW-1310M (Scion, Fredrickburg, VA). Images were enhanced using PhotoShop software. The sections were analyzed for the absolute number of modified pixels on histogram, using color replacer tool in Paint-Shop-Pro 7 software. The numerical results of the number of modified pixels were then transferred to excel for calculating average and standard deviation. The statistical significance of differences was recognized at p < 0.05 by using a t test.
The TUNEL (TdT-mediated dUTP Nick End Labeling) assay was performed according to the manufacturer's instructions using the In Situ Cell Death Kit, TMR red (Roche Diagnostics, Mannheim, Germany). Stained sections were viewed with Olympus CX-40 (Olympus Optical, Tokyo, Japan) and photographed using a gray scale digital camera CFW-1312M (Scion, Fredrickburg, VA). Images were enhanced using PhotoShop software.
RNA array analysis
All experiments were performed using Affymetrix Hu133A 2.0 oligonucleotide arrays, as described at (url1). Total RNA from each sample was used to prepare biotinylated target RNA, with minor modifications from the manufacturer's recommendations (url2). Briefly, 5 μg of mRNA was used to generate first-strand cDNA by using a T7-linked oligo(dT) primer. After second-strand synthesis, in vitro transcription was performed with biotinylated UTP and CTP (Affymetrix), resulting in ∼300-fold amplification of RNA. The target cDNA generated from each sample was processed as per manufacturer's recommendation using an Affymetrix GeneChip Instrument System (url2). Briefly, spike controls were added to 15 μg fragmented cRNA before overnight hybridization. Arrays were then washed and stained with streptavidin–phycoerythrin, before being scanned on an Affymetrix GeneChip scanner. A complete description of these procedures is available at (url2). Additionally, quality and amount of starting RNA was confirmed using an agarose gel. After scanning, array images were assessed by eye to confirm scanner alignment and the absence of significant bubbles or scratches on the chip surface. 3′/5′ ratios for GAPDH and beta-actin were confirmed to be within acceptable limits (0.96–1 and 1.2–1.28), and BioB spike controls were found to be present on all chips, with BioC, BioD and CreX also present in increasing intensity. When scaled to a target intensity of 150 (using Affymetrix MAS 5.0 array analysis software), scaling factors for all arrays were within acceptable limits (0.95–1.17), as were background, Q values and mean intensities. Details of quality control measures can be found at http://www.ncbi.nlm.nih.gov/geo/ or at http://eng.sheba.co.il/genomics.
The probe sets contained in the Affymetrix Human Hu133A2 oligonucleotide array or oligonucleotide arrays were filtered using Mas 5 algorithm. Treated and control samples were compared. The comparison generated a list of “active genes” representing probe sets changed by at least 2-fold (as calculated from the MAS 5 log ratio values) (LR ≥ 1 or LR ≤ −1) and detected as “increased” or as “decrease” (I or D, p = 0.0025) or marginal “increased” or as marginal decrease (MI or MD, p = 0.003) in all treated sample as compared with all the control samples in at least one time point. This list excluded up-regulated genes in all treated samples with signals lower than 20 or detected as absent, and down-regulated gene with baseline signals lower than 20 and detected as absent in the control samples. Hierarchical clustering was performed using Spotfire DecisionSite for Functional Genomics (Somerville, MA). Genes were classified into functional groups using the GO annotation tool.13 Overrepresentation calculations were done using Ease Douglas A.14 Functional classifications with an “ease score” lower than 0.05 were marked as overrepresented.
Semi-quantitative RT-PCR analysis
Whole cell RNA was extracted from xenografts using TRI Reagent (Molecular Research Center, Inc, Aurora, OH) following the manufacturer's instructions. Next, 0.5 μg of total RNA was reverse transcribed using the AccuPower RT/PCR Premix (Bioneer, Alameda, USA), and then amplified with the following primer pairs all from human:
NFkB p65 mRNA: forward primer 5′ TACACAGGACCAGG GACAGTGC 3′ and reverse primer 5′ AGCTGCCAG AGTTTCGGTTCAC 3′ (21 cycles).
NFkB p50 mRNA: forward primer 5′ TTGGTAGTGGCG GTGGAGGAG 3′ and reverse primer 5′ GTCTCCTGTCAC CGCGTAGTCG 3′ (21 cycles).
Interleukin (IL)-1β mRNA: forward primer 5′ ACAACAGGCTGCTCTGGGATTC 3′ and reverse primer 5′ GTGG TGGTCGGAGATTCGTAGC 3′ (22 cycles).
GAPDH mRNA: forward primer 5′ AAGGTCATCCCT GAGCTGAACG 3′ and reverse primer 5′ CAAAGGTG GAGGAGTGGGTGTC 3′ (21 cycles).
Laminin γ2 mRNA: forward primer 5′ GCCTCTGCTTCTCGCTCCTC 3′ and reverse primer 5′ AGCGGTCCCTTT CTCTGTGC 3′ (18 cycles).
Fibronectin mRNA: forward primer 5′ CACCTTCTTGGAGGCGACAAC 3′ and reverse primer 5′ AAACCTCGG CTTCCTCCATAAC 3′ (18 cycles).
IL8 mRNA: forward primer 5′ TCTTGGCAGCCTTCCT GATTTC 3′ and reverse primer 5′ CAACCCTCTGCACC CAGTTTTC 3′ (22 cycles).
Laminin α3 mRNA: forward primer 5′ TCTCCATGTGCCT CATCTGCTC 3′ and reverse primer 5′ AGCTGTCATCC CCTGACTGGAG 3′ (18 cycles).
Laminin β3 mRNA: forward primer 5′ GTCTTACCGA AGTCTGAGGAGC 3′ and reverse primer 5′ GGCAGATG AAATGCTGCAAGTG 3′ (18 cycles).
The selected primers were derived from separated exons. PCR was performed under linear conditions, which were found to be optimal for quantitative comparison of the xenograft mRNA levels. The expected PCR products were specific and could be differentiated from contamination by genomic DNA PCR products. 18S ribosomal RNA primer pairs (Quantum RNA, Ambion, Austin, TX) were used to yield a 489-bp fragment derived from the 18S rRNA as an internal control. PCR products were separated on a 1.4% agarose gel, and were visualized with ethidium bromide.
For siRNA transfection, the clones (1.6 × 105) were plated in 6-well plates, 1 day before transfection. The cells were transfected using the LipofectAMINE 2000 reagent (Invitrogen, Paisley, UK). Each transfection was performed by mixing 5 μL of 20 μM siRNA duplex with 5 μL of LipofectAMINE 2000 and 245 μL serum-free Opti-MEM (Gibco, Invitrogen, Paisley, UK) medium. The mixture was left for 20 min at room temperature and then was added to the cells. Cells were incubated for 8 hr, then the medium was exchanged with 4 mL fresh RPMI containing 10% FCS. The transfected cells were grown in their growth medium for an additional 24 hr, which was then replaced with HBSS (Beit Haemek, Israel) for 18 hr; thereafter the cells were harvested.
Synthetic small interfering RNA (siRNA) was designed for targeting sequences in the open reading frame of the human TMF mRNA. siRNA corresponding to nucleotides 64 to 83 of the human TMF mRNA open reading frame was designed (5′ AAGU CUAUUGACAGGGUUCUG 3′). This siRNA duplex was synthesized and purified by Dharmacon (USA). siRNA corresponding to nucleotides 706 to 725 of the human TMF mRNA open reading frame (5′ GGCAGAGCAAUACACCUUCTT 3′) was planned by Ambion Software Silencer predesigned siRNA (ID: 139703) and used for targeting the human TMF mRNA. This siRNA duplex was synthesized and purified by Ambion. Selected siRNA target sequences were also submitted to a BLAST search against other murine genome sequences to ensure target specificity. As a nonrelevant siRNA negative control, a sequence targeting the firefly (Photinus pyralis) luciferase (luc siRNA) (accession no. X65324) gene was used.
In vivo ubiquitination assay
Cells from established clones were cotransfected with 8 μg Flag-mp65-6×His and 2 μg Myc-Ubiquitin or 8 μg Flag-mp65-6×His and 2 μg pcDNA3 or 8 μg pcDNA3 and 2 μg Myc-Ubiquitin constructs. After transfection, the clones were incubated for 24 hr and then transferred to HBSS for 18 hr. The cells were then extracted in lysis buffer: 20 mM Tris, pH 7.5, 150 mM NaCl, 1% Nonidet P-40, 0.5% sodium deoxycholate, 2 mM Na2VO4, complete mini protease inhibitor (Roche Diagnostics, Mannheim, Germany) and 1 μL of nuclease, and rotated for 15 min at 4°C. The proteins were cleared by centrifugation at 14,000g for 7 min at 4°C. The pellet was solubilized with 1 mL guanidinium-HCl buffer (6 M guanidinium-HCl, 0.1 M Na2HPO4/NaH2PO4, pH 8.0). Protein levels were quantitated, and extracts were then mixed with 0.1 mL of metal (Co2+)-affinity Resin (BD Bioscience, Bedford, MA) for 4 hr at room temperature, in a vertically rotating 2-mL tube.15 The beads were successively washed with the following buffers: 1 mL of 6 M guanidinium-HCl buffer pH 8.0; 1 mL of 6 M guanidinium-HCl pH 5.8; 1 ml of 1:1 6 M guanidinium-HCl: protein buffer (0.1 M Na2HPO4/NaH2PO4 pH 8.0, 100 mM KCl, 20% glycerol and 0.2% NP-40); 1 mL of 1:3 6 M guanidinium-HCl: protein buffer; and 1 mL of protein buffer. Elution was performed with 30 μL Laemmli buffer containing 10 mM EDTA. Eluted proteins were resolved in 10% SDS-PAGE for Western blot analysis.
Ectopic expression of TMF/ARA160 does not affect the cell-cycle progression of PC3 cells, but attenuates the growth of PC3 xenografts
To examine whether TMF could impair the growth of solid tumors in vivo, and to identify cellular regulators that are governed by TMF, the protein was ectopically expressed in human PC3 prostate cancer cells, which do not express Stat3 (Fig. 1a). These cells were stably transfected with a pcDNA3-based vector encoding either the intact, HA-tagged TMF, or an HA-TMF mutant devoid of the functional “BC-box” motif.8 Cells transfected with an empty vector served as a control. Clones expressing the HA-tagged TMF proteins were selected (Fig. 1a) and their subcellular distribution was determined. Like the endogenous protein, the transfected TMF was associated with the Golgi. However, the localization of the HA-TMF/ARA160 was not confined to the Golgi alone and a fraction of the ectopic proteins could be detected in the cytoplasm (Fig. 1b).
To determine whether ectopic TMF affects the proliferation and/or survival of PC3 cells, selected clones were subjected to flow-cytometry analysis and a proliferation assay (MTT). In both assays, no significant difference was seen between the HA-TMF transfected and control clones (Fig. 1c and 1d). Thus, under normal growth conditions TMF does not affect the proliferation or survival of PC3 cells.
To examine the effect of TMF on tumor growth, PC3 clones were subcutaneously injected to athymic “nude” mice, and the development of xenografts was followed. The development of xenografts expressing ectopic HA-TMF was significantly impaired in comparison to PC3 xenografts that express the mutated TMF or which harbor the expression vector alone (Fig. 2).
TMF/ARA160 impairs angiogenesis and imposes apoptotic death in PC3 xenografts
The profound attenuated progression of the xenografts expressing ectopic HA-TMF, prompted us to investigate whether this reflects impaired proliferation and/or survival of the expressing tumor cells. Sections were prepared from the tumors 25 days after inoculation (Fig. 3a and 3b) and were stained for proliferation and cell death–associated markers. Although the percentage of cells expressing the proliferation marker PCNA was similar in the control and in HA-TMF-expressing xenografts (Fig. 3c), the level of activated caspase 3, which reflects the onset of apoptotic death, was significantly higher in sections prepared from HA-TMF-expressing xenografts (Fig. 3d). Accordingly, the number of cells exhibiting DNA fragmentation as reflected by the TUNEL assay positive spots was significantly higher in HA-TMF sections (Fig. 3e).
Various factors could lead to the onset of apoptotic death in solid tumors. One of these could be impaired angiogenesis resulting in impaired nutrient exchange in the developing solid tumor.16 To estimate the relative blood vessel mass in the HA-TMF-expressing and nonexpressing xenografts, sections prepared from the 2 tumor types were stained with antibodies directed toward the endothelial cell marker, CD31. This staining revealed a significantly reduced level of blood vessels in the HA-TMF-expressing xenografts, compared with the control xenografts (Fig. 3f).
TMF/ARA160 downregulates proangiogenic genes
TMF was shown to recruit a key transcription factor, Stat3, to proteasomal degradation.8 To examine whether ectopic TMF affects the expression of genes during xenograft progression, the RNA expression profiles of 22-day tumors, from control and HA-TMF-expressing xenografts, were compared using a RNA-array analysis. Emphasis was placed on genes involved in the regulation of apoptosis and angiogenesis. Although the expression of both the antiapoptotic gene bclx,17 and the proapoptotic gene bik,18–20 were downregulated in HA-TMF-expressing xenografts (data not shown), a rigorous decreased expression of proangiogenic genes was seen in HA-TMF-expressing tumors (Fig. 4a). The downregulation of 5 selected proangiogenic genes was verified using a semiquantitative RT-PCR analysis (Fig. 4b). These included the genes encoding for laminin alpha 3, laminin beta 3, laminin gamma 2, interleukin 8 and interleukin 1β, all exerting a proangiogenic activity.21–23 The levels of the RNAs encoding IL-8 and IL-1β were further analyzed and found to be decreased in 3 independent HA-TMF-expressing xenografts (Fig. 4c). Thus, TMF preferentially downregulates proangiogenic genes, and this could underlie the inhibition of angiogenesis mediated by TMF in the developing xenografts.
TMF/ARA160 downregulates the accumulation of p65/RelA in metabolically stressed PC3 clones
To gain insight into mechanisms which mediate the inhibitory effect of TMF on proangiogenic genes, we sought to examine the effect of TMF on transcription factors that regulate the angiogenic process. The genes encoding IL8 and IL-1β are controlled by the transcription factor NFκB,24 which is comprised of 2 main components—p65/RelA and p50.25 To examine whether TMF affects the cellular accumulation of p65 and/or p50, the level of the 2 proteins was determined in xenografts expressing the ectopic HA-TMF, and in controls. Although the levels of p50 and a nonrelated transcription factor-AP2α were marginally reduced in HA-TMF-expressing tumors, the level of p65/RelA protein was decreased by ∼2-fold in xenografts expressing HA-TMF, as shown by Western blot (Fig. 5a). Comparison of the p65 and p50 mRNA levels in the 2 tumor types, using a semiquantitative RT-PCR, did not reveal a significant decrease in the RNAs encoding p65 or p50 (Fig. 5b).
To further characterize the mechanism by which TMF affects the cellular accumulation of p65, we used HA-TMF-expressing and nonexpressing PC3 clones grown in vitro. Because tumor cells experience metabolic constraints in vivo,1 the level of p65 was determined in PC3 clones, grown under normal or conditions of serum, and nutrient deprivation. The level of p65/RelA in both soluble and insoluble fractions was determined in all cells analyzed. Although under normal growth conditions, the level of the soluble p65 was only slightly higher in cells expressing either native or “BC-box” mutated HA-TMF, serum and nutrient starvation induced a different effect in the 3 tested clone types. Extended starvation led to a decrease in the level of p65 in control clone and in the native HA-TMF-expressing cells, but not in the B15 clone, which expresses the “BC-box” mutated HA-TMF (Fig. 6a). However, while in the control clones, the proteasome inhibitor MG132 restored the level of the soluble p65, in HA-TMF-expressing cells, treatment with the proteasome inhibitor did not reverse this effect (Fig. 6a). Starvation did increase the level of insoluble p65 in HA-TMF-expressing cells and furthermore, treatment with MG132 restored the level of p65 in the insoluble fraction of HA-TMF-expressing clones (Fig. 6b). Thus, TMF directs the proteasomal degradation of p65/RelA under serum and nutrient deprivation, and interference with the proteasome activity leads to the accumulation of an insoluble26 p65/RelA fraction in HA-TMF-overexpressing cells. This effect of TMF on p65 was not seen when PC3 clones were subjected to hypoxic conditions, which could also resemble a metabolic burden imposed on solid tumor cells (data not shown). Notably, starvation increased the level of the exogenous TMF in both the soluble and insoluble fractions of the PC3 clones (Fig. 6a and 6b).
To corroborate the involvement of TMF in the downregulation of p65 under serum and nutrients withdrawal, a complementary approach was applied. TMF was knocked down in starved PC3 clones, using a specific siRNA directed toward the TMF mRNA. Reduction of TMF increased the level of p65, in starved PC3 clones (Fig. 7), thus indicating the involvement of TMF in the downregulation of p65 under those conditions.
TMF/ARA160 directs the ubiquitination of p65/RelA in metabolically stressed PC3 clones
To examine whether ubiquitination of p65 could mediate the downregulation of this protein by TMF, a TMF driven in vivo ubiquitination assay was applied. P65 was tagged with 2 tags: Flag and the 6×His sequence. HA-TMF-expressing and nonexpressing clones were cotransfected with a plasmid encoding the tagged p65 and a plasmid encoding a Myc-tagged ubiquitin. Because the TMF-directed ubiquitination and proteasomal degradation depends on the integrity of the TMF “BC-box”,8 a B15 clone, which bears a “BC box” deficient TMF, was transfected as well. Clones were then starved for 18 hr and were also treated with the proteasome inhibitor, MG132, 4 hr before harvest. Cell extracts were prepared and then divided into soluble and insoluble fractions. Expression of the tagged p65 protein and the Myc-ubiquitin induced the accumulation of insoluble ubiquitinated proteins in a clone expressing the intact HA-TMF, but not in a clone expressing a HA-TMF mutant devoid of the “BC box” element (B15), nor in a clone which did not express HA-TMF (Fig. 8a and b). To examine whether p65 becomes ubiquitinated in the HA-TMF-expressing clone, the insoluble fraction was recovered in a denaturing guanidine buffer. Denatured Flag-His-tagged p65 protein was affinity purified on a cobalt-Sepharose resin. This approach removed proteins which noncovalently associated with His-tagged p65. The purified p65 was resolved in SDS-PAGE and was then reacted with anti-6×His and anti-Myc antibodies. Our results clearly demonstrated the ubiquitination of p65/RelA in HA-TMF-expressing cells (Fig. 8c), thus implying the TMF driven ubiquitination of p65 and corroborating the regulatory effect of TMF on this subcomponent of NFκB.
Mammalian cells respond in various ways to stress insults elicited in them. While in some cases, cells respond by an attempt to overcome the effects of stress, under other circumstances, stressed cells are designated for programmed cell death (apoptosis).27 Exploration of the molecular mechanisms that underlie these processes revealed the involvement of intracellular organelles in the modulation of cellular responses to imposed stress. The most profound example of an organelle-directed response to cellular stress is the release of cytochrome C and the apoptosis-inducing factor from the mitochondria of cells that experience genotoxic burden.28–30
In the current study, we demonstrate the involvement of a protein associated with the Golgi apparatus, TMF/ARA160, in the response of cells to metabolic stress insult. Overexpression of TMF did not affect the proliferation or survival of PC3 prostate carcinoma clones, grown in culture under normal growth factor and nutrient supply. However, overexpressed TMF did attenuate the progression of PC3 xenografts in vivo. This resulted mainly from impaired angiogenesis and the onset of apoptotic death in the developing tumors. In accordance with that notion the more profound effect of TMF was seen in late stages of tumor development.
The preferential proapoptotic effect of TMF could be directly linked to its antiangiogenic activity. Alternatively, it may reflect the combined effect of TMF on proangiogenic genes on one hand, and on pro or antiapoptotic genes on the other. It should be noted that TMF decreased the expression of both a proapoptotic gene-bik,18–20 and the bcl-xL RNA, which encodes for an antiapoptotic protein-Bcl-XL17 (data not shown).
An interrupted angiogenic process results in metabolic stress, which can lead to apoptotic death16 in tumors, such as PC3 xenografts, which harbor a functional apoptotic machinery.31, 32 The proliferation rate of cells in surviving areas of the HA-TMF-expressing xenografts did not differ from that detected in nonexpressing tumors. This is consistent with the lack of effect of ectopic TMF on the cell-cycle profile of PC3 cells, in culture. Notably, ectopic TMF also did not induce apoptotic death in PC3 clones grown under normal conditions in culture. Therefore, it seems that TMF exerts its growth inhibitory effect only under metabolic constraints which exist in developing solid tumors in vivo. Specifically, TMF seems to become active under nutrient deprivation rather then under oxygen shortage which leads to hypoxic stress.
To gain insight to the mechanisms that underlie this effect of TMF, comparative analysis of RNA expression profiles was performed. Our results indicated the downregulation of several proangiogenic genes. The suppressive effect on proangiogenic genes, most probably explains the antiangiogenic activity of TMF. Interestingly, the level of the vegf mRNA was not decreased in HA-TMF xenografts (data not shown). This is in compliance with the notion that degradation of Stat3, which activates the vegf gene,33–35 does not mediate the antiangiogenic activity of TMF in PC3 cells. Among the proangiogenic genes downregulated by TMF were also the genes encoding for IL-8 and IL-1β. These genes are positively controlled by the transcription factor NFκB,24 thereby mediating the proangiogenic activity36 of this key transcription factor. We show here that TMF directs the ubiquitination and proteosomal degradation of the NFκB subunit, p65/RelA,25 under metabolic stress conditions. Several regulatory pathways were shown to direct the degradation of p65/RelA and to terminate NFκB activity.37–39 However, the degradation of p65/RelA under stress conditions has not been reported. In the current study, we demonstrate the TMF-driven downregulation of p65/RelA, under metabolic stress. Although some stress cues, such as ionizing radiation, activate NFκB,40 metabolic stress downregulates this transcription factor. TMF is, therefore, part of a response system which determines the destiny of cells exposed to an extended metabolic constraint. The metabolic limitations experienced by expanding solid tumors2 could lead to the activation of the TMF/ARA160 system, resulting in the downregulation of the key regulator, NFκB. Due to the proangiogenic41 and antiapoptotic activities of NFκB,42–45 which support tumor progression, its down regulation would be expected to significantly attenuate tumor growth. The effect of TMF on NKκB and the tumor attenuating activity of this Golgi-associated protein, could explain the observed decrease in the TMF level in malignant brain tumors.8 The downregulation of TMF would enable expanding tumors to circumvent its suppressive effects on angiogenesis. Elucidating the mechanisms through which solid tumors handle the suppressive activity of TMF may open new avenues for manipulating the TMF system, and thereby may enable novel approaches for restraining tumor progression.