Fax: (706) 721-6608
Cyclic 3′,5′-guanosine monophosphate-dependent protein kinase inhibits colon cancer cell adaptation to hypoxia
Version of Record online: 11 MAY 2011
Copyright © 2011 American Cancer Society
Volume 117, Issue 23, pages 5282–5293, 1 December 2011
How to Cite
Kwon, I.-K., Wang, R., Prakash, N., Bozard, R., Baudino, T. A., Liu, K., Thangaraju, M., Dong, Z. and Browning, D. D. (2011), Cyclic 3′,5′-guanosine monophosphate-dependent protein kinase inhibits colon cancer cell adaptation to hypoxia. Cancer, 117: 5282–5293. doi: 10.1002/cncr.26192
- Issue online: 18 NOV 2011
- Version of Record online: 11 MAY 2011
- Manuscript Accepted: 14 MAR 2011
- Manuscript Revised: 8 MAR 2011
- Manuscript Received: 24 OCT 2010
- cyclic 3′;5′-guanosine monophosphate;
Type 1 cyclic 3′,5′-guanosine monophosphate-dependent protein kinase (PKG) has recently been reported to inhibit tumor growth and angiogenesis. These effects suggest that PKG activation may have therapeutic value for colon cancer treatment, but the signaling downstream of this enzyme is poorly understood. The present study examined the mechanism underlying the inhibition of angiogenesis by PKG.
The effect of ectopically expressed PKG on colon cancer cell adaptation to a 1% O2 (hypoxic) environment was examined in vitro by measuring hypoxic markers, cell death/viability, and hypoxia inducible factor (HIF) activity.
Ectopic PKG inhibited angiogenesis in SW620 xenografts and significantly attenuated hypoxia-induced increases in vascular endothelial growth factor at both the mRNA and protein levels. PKG activation also blocked hypoxia-induced hexokinase 2 expression, which corresponded with reduced cellular adenosine triphosphate levels. Moreover, PKG expression significantly reduced cell viability and promoted necrotic cell death after 2 days in a hypoxic environment. To gain some mechanistic insight, the effect of PKG on HIF activation was determined using luciferase reporter assays. PKG activation inhibited HIF transcriptional activity in several colon cancer cell lines, including SW620, HCT116, and HT29. The mechanism by which PKG can inhibit HIF activity is not known, but it does not affect HIF-1α protein accumulation or nuclear translocation.
These findings demonstrate for the first time that PKG can block the adaptation of colon cancer cells to hypoxia and highlights this enzyme for further evaluation as a potential target for colon cancer treatment. Cancer 2011;. © 2011 American Cancer Society.
Several disparate lines of investigation have converged on the cyclic 3′,5′-guanosine monophosphate (cGMP) signaling axis as having therapeutic potential for the treatment or prevention of colon cancer. 1, 2 One line of study identified enterotoxigenic Escherichia coli as a likely cause of the lower prevalence of colorectal cancer in developing countries compared with industrialized ones. 3, 4 The mechanism is thought to involve a heat-stable toxin produced by the bacteria that mimics endogenous uroguanylin. This peptide ligand binds to receptor guanylyl cyclase C on the intestinal epithelium, causing increased cGMP levels, and can suppress tumor burden in the ApcMin/+ mouse cancer model. 5, 6 Signaling through guanylyl cyclase C has tumor-suppressive properties in the intestine, where it inhibits proliferation along the crypt-villus axis. 7, 8
The nature of the tumor-suppressive properties of uroguanylin is presently not known, but as central mediators of cGMP signaling in cells, cGMP-dependent protein kinase (PKG) is a likely mediator. 9-11 Evidence for growth-inhibitory effects of PKG comes from studies with exisulind, which increases cGMP levels by inhibiting phosphodiesterases. 12-14 This drug can induce apoptosis in some colon cancer cells in a type 1 PKG-dependent manner. 12, 15, 16 The signal transduction pathways involved in the antitumor effects of PKG are not understood, but in vitro studies have identified activation of cJun-N-terminal kinase, 17 down-regulation of β-catenin/T-cell factor activity, 12, 18 and activation of SP1. 19 More recently it was reported that PKG can inhibit β-catenin/T-cell factor in colon cancer cells by activation of FoxO4, which effectively competes with T-cell factor for β-catenin. 20 However, this pathway did not significantly affect the growth or apoptosis of metastatic SW620 colon cancer cells grown in vitro, 21 but potently slowed growth and caused extensive cell death in xenografts. 22 The contrast in effect of PKG in vitro and in vivo is most likely because of the ability to block tumor angiogenesis. 23
The present study has addressed the mechanism underlying the antiangiogenesis effect, and it was found that ectopic PKG can inhibit hypoxia inducible factor (HIF)-dependent transcription. This in turn results in less vascular endothelial growth factor (VEGF) synthesis and reduced metabolic adaptation to the hypoxic tumor microenvironment. These findings underscore the potential for PKG as part of a therapeutic strategy to combat colon cancer.
MATERIALS AND METHODS
Tissue Culture and Reagents
All cell lines were obtained from the American Type Culture Collection (ATCC; Manassas Va) and maintained in 5% CO2 in RPMI-1640 medium containing 10% fetal bovine serum (FBS), and supplemented with 200 μM L-glutamine, 10 IU/mL penicillin, and 10 mg/mL streptomycin. In experiments comparing normoxia (20% O2) and hypoxia (0.5%-1% O2), cells were incubated in a glove-box hypoxic chamber (Coy Laboratory Products, Grass Lake, Mich). We have previously described the creation and characterization of SW620 colon carcinoma cell lines made inducible for type 1 PKG expression in response to mifepristone. The medium used to maintain stocks of the H3Z6 clone of inducible SW620 was supplemented with 300 μg/mL each of hygromycin and Zeocin. Before experimentation, the cells were grown up for at least 1 passage in the absence of antibiotics. The mifepristone, 4,6-diamidino-2-phenylindole (DAPI), and 8Br-cGMP were from Calbiochem (San Diego, Calif). NP-40 and Tween-20 were from Sigma (St Louis, Mo), and all other chemicals were from Fisher Scientific (Pittsburgh, Pa).
Antibodies and Constructs
The antibody specific for hexokinase 2 (HK2) was from Cell Signaling (Beverly, Mass), whereas antibodies for HIF-1α and β-actin were from Sigma. The polyclonal anti-PKG1 antibodies raised against the common C-terminus, and the PKG expression vectors, have been described previously. 24 The HIF-responsive luciferase reporter was a generous gift from Dr. William Kaelin (Dana-Farber Cancer Institute).
Angiogenesis in Xenografts
Subcutaneous tumor studies were carried out using engineered SW620 cells in nude mice (Harlan Laboratories, Indianapolis, Ind) essentially of as described previously. 22 The perfused tumor vasculature was observed using Microfil MV-122 (Flow Tech, Carver, Mass) perfusion as described in detail elsewhere. 25, 26 Briefly, the tumor-bearing mice were anesthetized, and heparin was delivered through a 27-gauge needle into the left ventricle, followed by Microfil. After overnight curing at 4°C, the tumors were excised, fixed in formalin, and sectioned on a vibratome (50-100 μm). The tumor blood vessels were imaged and quantified as mean vessel density and mean vessel size (length × diameter) using Scion Image software (Scion Corporation, Frederick, Md). Mean densities were obtained for at least 5 sections of the tumor.
Cells were grown on coverslips in 6-well dishes until 50% confluent and then treated with 100 μM 8Br-cGMP and 1 nM mifepristone (PKG inducer) and either maintained in normoxia or transferred to the hypoxic chamber. At intervals the cells were washed with phosphate-buffered saline (PBS), fixed in 4% paraformaldehyde, and permeabilized with 0.1% Triton-X-100. After blocking for 1 hour at 37°C in PBS containing 5% goat serum, the coverslips were incubated overnight at 4°C in blocking buffer containing a 1:100 dilution of anti-HIF-1α (Sigma). Coverslips were then incubated for 2 hours at room temperature with Alexa Fluor 568 goat antimouse secondary antibodies (Invitrogen, Carlsbad, Calif) and counterstained with DAPI for 5 minutes, before affixing to slides using Mowiol (Calbiochem, San Diego, Calif). Images were captured using a Nikon (Melville, NY) TE-300 inverted epifluorescence microscope equipped with a SPOT RT3 camera and SPOT Software version 4.7 (Diagnostic Instruments, Sterling Heights, Mich).
Cells were lysed by incubation with ice cold lysis buffer (50 mM HEPES pH 8.0, 150 mM NaCl, 1% Nonidet P-40, 0.25% deoxycholate) supplemented with protease inhibitor cocktail (Calbiochem, San Diego, Calif). Lysates were clarified by centrifugation, boiled in polyacrylamide gel electrophoresis buffer, and separated on 10% mini-gels followed by electrophoretic transfer to nitrocellulose. The blots were blocked with 5% bovine serum albumin in PBS containing Tween 20 and incubated with primary antibodies overnight at 4°C. After addition of 1:3000 peroxidase-conjugated secondary antibody (Bio-Rad Laboratories, Hercules, Calif), the bands were observed using chemiluminescence according to the manufacturer's instructions (Pierce, Rockford, Ill).
Measurement of VEGF Expression
The secretion of VEGF by colon cancer cells was performed using a human VEGF enzyme-linked immunosorbent assay (ELISA) kit according to the manufacturer's instructions (R&D Systems, Minneapolis, Minn). For these experiments, 2 × 105 cells in 5% FBS medium were seeded into triplicate wells of a 6-well dish and incubated 24 hours in normoxic or hypoxic conditions. The growth medium was then cleared by centrifugation and used directly in the assay.
Semiquantitative reverse transcriptase (RT) polymerase chain reaction (PCR) was used to determine relative VEGF gene expression levels. Total RNA was isolated using TRIzol reagent according to the manufacturer's instructions (Invitrogen), and cDNA generated using the GeneAmp RNA PCR kit (Applied Biosystems, Foster City, Calif). PCR was performed using 1 μL RT product in reactions with 0.2 U TaKAra Taq (Fisher Scientific) for 25 cycles at 60°C anneal temperature. The primers used for VEGF gene amplification were forward: CCATCCTGT GTGCCCCTGATG, and reverse: TGCCTCGCCTTG CAACGCGA. These primers can amplify several VEGF-A variants, but a single major variant and a weaker minor variant were detected in SW620 cells that generally migrated as a single band (data not shown). Control reactions used primers specific for hypoxanthine guanine-phosphoribosyltransferase (HPRT1 forward: 5′-CCATCCTGTGTGCCCCTG ATG, and reverse: 5′-TGCCTCGCCTTGCAACGCGA.
For measurements of transcription, the cells were cultured in 12-well plates, and triplicate wells were transfected with luciferase-reporter plasmids using Lipofectamine 2000 reagent (Invitrogen). Measurement of HIF-transcriptional activity was carried out by cotransfecting cells with 0.05 μg hypoxia responsive element-luciferase reporter/well and 0.2 μg cytomegalovirus (CMV)-β-galactosidase to control for cell number and transfection efficiency. After 24 hours recovery, the cells were either untreated or stimulated with 100 μM 8Br-cGMP before incubating in either normoxic or hypoxic conditions for an additional 6 to 8 hours before enzyme assay. Cell extracts were then prepared and analyzed for luciferase and β-galactosidase activities as described previously. 23 The relative luciferase activity was determined by subtracting the background values from each activity and then dividing the corrected luciferase values by the corrected β-galactosidase activity.
Cell Viability Determination
Colony formation assays were performed to assess cell viability after intervals of exposure to hypoxia in the presence or absence of PKG. In these experiments, the cells were cultured in 12-well plates and treated for PKG expression and incubated under hypoxic conditions. After 48 hours, cells were washed with PBS and trypsinized, and equal volumes of the cell suspensions were transferred to 6-well dishes and allowed to form colonies over a period of 10 days. To quantify colony formation, the cells on the dishes were fixed in 100% methanol for 30 minutes and stained with KaryoMax Gimsa stain (Sigma) for 1 hour. The 6-well dishes were scanned at high resolution, and the total colony area for each well was determined using Scion Image software (Scion Corporation).
To evaluate direct counting of living and dead cells, trypan blue assay was performed (Mediatech, Manassas, Va). Cells were grown in 6-well dishes until 50% confluent and then either untreated or incubated with 100 μM 8Br-cGMP and 1 nM mifepristone, before culturing in normoxic or hypoxic conditions. After 48 hours, the cells were detached and resuspended in PBS. Equal volumes of 0.4% trypan blue dye were added, and after 5 minutes at room temperature the living (clear) and dead (blue) cells were counted by using a hemocytometer.
Measurement of Metabolic Indicators
To determine adenosine triphosphate (ATP) levels in cell extracts, the CellTiter-Glo luminescent cell viability assay was used essentially as recommended by the manufacturer's instructions (Promega Corp, Madison, Wis). Briefly, the cells (1 × 104 cells/mL) were treated with 100 μM 8Br-cGMP and 1 nM mifepristone and cultured in either normoxic or hypoxic conditions for 48 hours in 96-well plates (100 μL/well). Before assay, the plate was equilibrated for 30 minutes at room temperature, and then equal volume (100 μL) of CellTiter-Glo reagent was added to the cell culture medium present in each well. Before reading the luminescence, the plate was incubated for 10 minutes to stabilize the luminescent signal.
The 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) cell proliferation assay was used to measure mitochondrial reductases as an indicator of aerobic metabolism. For these studies, the cells were cultured in 96-well plates and at intervals after treatment MTT assays were performed according to the manufacturer's instructions (ATCC).
Cell Death Determination
Apoptotic or necrotic cell death was assessed using annexin-V fluorescein isothiocyanate (FITC) and propidium iodide double staining according to the manufacturer's instructions (Annexin-V FITC Apoptosis Kit; BD Bioscience, San Diego, Calif). H3Z6 cells (PKG inducible clone) were seeded in 12-well plates at 50% confluence and either untreated or treated with 100 μM 8Br-cGMP and 1 nM mifepristone. After treatment, the cells were then cultured for 48 hours in a normoxic or hypoxic conditions. Cells were washed twice in cold PBS, harvested by treatment with a trypsin-ethylenediaminetetraacetic acid solution, centrifuged, and resuspended in 100 μL annexin-V binding buffer. Then, 5 μL of FITC-labeled annexin-V and 5 μL of propidium iodide buffer was added to the resuspended cells and incubated for 15 minutes at room temperature in the dark. Immediately before analysis by flow cytometry, 400 μL annexin-V binding buffer was added to each tube.
All quantitative experiments were reproduced in at least 3 independent experiments with multiple wells in each replicate. The resulting quantitative data were expressed as means with error bars indicating standard error of the mean. Two means were considered to be statistically significant when 2-sided Student t test (Excel; Microsoft, Redmond, Wash) produced P values that were <.05.
Ectopic PKG1 Expression Blocks Classical Cellular Responses to Hypoxia in Colon Cancer Cells
It has been reported previously that ectopically expressed type 1 PKG causes reduced growth of colon cancer xenografts. 22, 23 The PKG-expressing tumors contained reduced levels of VEGF and decreased overall CD31 staining, suggesting that PKG inhibits angiogenesis. 23 The present study sought to extend this work by measuring vasculature in xenografts using a Microfil perfusion approach. 26 Confirming the previous findings, these experiments revealed significantly less perfused blood vessels in the tumors induced to express PKG compared with uninduced tumors or those derived from parental SW620 cells (Fig. 1A, B). To gain some insight into the antiangiogenic mechanism, the VEGF levels were measured in colon cancer cells grown in vitro. These studies revealed that ectopic PKG did not affect the basal levels of VEGF produced by the SW620 cells under standard tissue culture conditions, but significantly inhibited the increase in VEGF production by cells maintained in a hypoxic environment (Fig. 1C). This PKG-dependent blockade of VEGF production in colon cancer cells was also observed at the RNA level, which suggested that the inhibitory mechanism involved a fundamental regulation of hypoxia-induced gene expression (Fig. 1D). This notion prompted us to examine the well-characterized hypoxia-induced expression of glycolytic enzymes, which serves to increase anaerobic ATP production. 27, 28 There was a significant basal amount of HK2 in cells cultured under normoxic conditions, but the level increased when either the parental cells or uninduced SW620 clones were incubated under hypoxic conditions (Fig. 2A). However, in the presence of ectopic PKG, the increase under hypoxic conditions was not observed, and the HK2 remained at basal levels. Importantly, the PKG-expressing SW620 colon cancer xenografts also exhibited much lower HK2 levels compared with xenografts derived from either uninduced or parental cells (Fig. 2B). We also observed lower enolase levels in PKG-expressing colon cancer cells under hypoxic conditions, suggesting a general reduction in glycolytic capacity (data not shown). In support of this idea, the cellular ATP levels were significantly lower in the PKG-expressing cells compared with the uninduced controls under hypoxic conditions (Fig. 2C). The lower ATP levels measured in PKG-expressing cells during hypoxia were not because of defective aerobic metabolism, because there was no effect of PKG on reducing activity measured by MTT assay (Fig. 2D).
PKG1 Induces Necrotic Cell Death in Colon Cancer Cells Under Hypoxic Conditions
Because glycolytic enzymes and respective ATP levels were found to be lower in PKG-expressing cells under hypoxic conditions, it was important to determine whether this effect of PKG compromised cell viability. To test this, SW620 colon cancer cells were maintained in a hypoxic environment, and at intervals the viable cells were quantified by colony formation assays (Fig. 3). Relative to normoxia, there was a significant reduction in cell viability in all cells incubated under hypoxic conditions, but this effect was significantly more pronounced in PKG-expressing cells than in uninduced controls after 48 hours. Cell proliferation was arrested under the severe hypoxic conditions irrespective of PKG expression, and the total cell number did not differ significantly between control and PKG-expressing cells after 48 hours in hypoxia, but many of the latter cells were no longer adherent and were floating in the medium (data not shown). This observation was confirmed by vital dye staining, which showed a 2-fold higher (47% of total) trypan-positive count in cells induced to express PKG (Fig. 3C). The cell death detected by trypan blue staining was not associated with poly adenosine diphosphate-ribose polymerase cleavage, indicating that apoptosis was not part of the mechanism (Fig. 3D). To gain further insight into the mechanism of cell death induced by PKG during hypoxia, we used flow cytometry to measure annexin binding and permeability to propidium iodide. These experiments demonstrated significant elevations in double-positive cells and in propidium staining alone, but there was no difference in the annexin staining alone between the different groups (Fig. 4). These results do not support the idea that PKG initiates apoptosis, but instead demonstrate that it promotes necrotic cell death under hypoxic conditions.
PKG Inhibits HIF Transcriptional Activity Without Affecting Expression or Localization
The inability of colon cancer cells to up-regulate either VEGF or glycolytic genes when expressing active PKG suggests that this enzyme exerts its effect at a fundamental stage of the hypoxia response. The central transcription activators that mediate the hypoxia-associated transcriptome are the HIFs. These proteins are usually subject to proteasomal degradation, but under low oxygen tension they accumulate and can activate transcription of a plethora of genes containing hypoxia-responsive elements in the promoter. 29, 30 To determine whether PKG could interfere with HIF activity, several colon cancer cell lines were cotransfected with HIF-luciferase reporter plasmids and either empty vector or PKG (Fig. 5A). Under hypoxic conditions, the basal luciferase activity increased relative to normoxia, but there was considerable variation in the hypoxic response in different cells types. The increase in HIF-reporter activity after 24 hours in hypoxia ranged from a 10-fold increase (SW620 cells) to a 2-fold increase (SW480 cells), but PKG was able to significantly block HIF transcription in all cells examined. These results indicate that the PKG-dependent inhibition of the hypoxia response is widespread, and not restricted to clonal or cell type-specific artifacts. The most dramatic effect of PKG was in the HCT116 cells, where the hypoxia-induced HIF activity was almost completely abrogated. These cells were therefore used to further show that the inhibition of HIF was not an overexpression artifact, as significant inhibition was observed at PKG levels that were undetectable by Western blot. Moreover, the inhibition of HIF was specific to PKG activity, because neither PKG expression alone nor treatment with cGMP was effective by itself (Fig. 5B, C). The regulation of HIF activity by PKG has not previously been reported, and owing to the significance of this pathway to cancer cells, the inhibitory mechanism used is potentially important. The principal process mediating HIF regulation is increased HIF protein expression in response to reduced hydroxylation in hypoxia. However, our preliminary experiments using inducible SW620 colon cancer cells did not detect an effect of PKG activity on the increase in HIF-1α expression (Fig. 5D). To determine whether PKG might prevent nuclear translocation of the HIF, we also looked at endogenous HIF proteins in control or PKG-expressing colon cancer cells using immunocytochemistry (Fig. 6). Results from these experiments were consistent with the Western blot data that showed that PKG expression did not alter the hypoxia-induced increase in HIF protein. Moreover, the majority of the HIF protein exhibited nuclear localization under hypoxic conditions, and this also was not significantly affected by PKG. Similar results were obtained in HCT116 cells transiently transfected to express PKG (data not shown).
Type 1 PKG expression is reduced in colon cancer relative to normal adjacent tissue, and re-expression in SW620 colon cancer cells has been reported to produce smaller xenografts with extensive necrotic regions. 22 It was subsequently found that the PKG-expressing tumors produced less VEGF than PKG-deficient controls, and exhibited defective angiogenesis. 23 The present study has demonstrated that ectopic PKG can inhibit the hypoxia response at the level of HIF transcription activation. As a well-characterized HIF target gene, the reduced VEGF levels and resulting defect in angiogenesis are explained by this effect of PKG. In addition to increased VEGF expression, cancer cells adapt to hypoxic tumor microenvironments by increasing glycolytic enzymes to compensate for loss of aerobic metabolism. 28, 29, 31 Consistent with the notion that PKG blocks adaptation to hypoxia at an early stage, the increase in glycolytic enzymes was also blocked in PKG-expressing cells, and this was associated with lower ATP levels and reduced viability under hypoxic conditions. Necrotic cell death in the absence of apoptosis is a hallmark of decreased cellular energy levels, and explains previous reports of extensive necrotic regions in xenografts expressing PKG. Results shown here are consistent with previous reports demonstrating that the cGMP signaling axis does not induce apoptosis in colon cancer cells. 21, 23
In light of the present findings, the defective angiogenesis first identified in SW620 xenografts expressing exogenous PKG was merely a symptom of a more fundamental effect of this enzyme on colon cancer cells. However, angiogenesis is among the hallmarks of cancer, and starving tumors by blocking this process is a well-recognized therapeutic strategy. 32, 33 A problem with antiangiogenic therapy is the development of resistance, 34, 35 and it has been suggested that the increased hypoxic tumor microenvironment created by antiangiogenic agents is selective for metastatic progression. 36-38 The ability of PKG to block adaptation to hypoxia underscores the potential of PKG as part of a therapeutic approach to combat colon cancer, possibly in combination with antiangiogenic agents.
Because PKG did not affect hypoxia-induced HIF expression or its translocation to the nucleus, it is likely that the effect of PKG is on the transcription activity of HIF itself. HIF regulation can be complex, and several mechanisms have been described for expression-independent inhibition of HIF activity, including sumoylation 39 and phosphorylation. 40 Another possibility is competition for β-catenin by FoxO4 activation, which has recently been shown to mediate inhibition of T-cell factor signaling by PKG in colon cancer cells. 20 Because β-catenin has been shown to enhance adaptation to hypoxia by directly binding to HIF-1α, 41, 42 it is possible that regulation of β-catenin availability by PKG could contribute to HIF inhibition in colon cancer cells, but this remains to be investigated.
In summary, the present work has identified a novel PKG-dependent pathway in colon cancer cells that can block adaptation to hypoxic tumor microenvironments. The inhibitory effect is at the level of HIF activity, but more work will be required to understand the mechanism. This finding explains previous reports of antitumor properties of PKG and establishes this enzyme as a potential therapeutic target for colon cancer.
The work was supported by American Cancer Society grant #RSG-07-174-01-CSM and MCG Research Institute grant #PSRP-00,037 to D.D.B.
CONFLICT OF INTEREST DISCLOSURES
The authors made no disclosures.