Phosphodiesterase-4 promotes proliferation and ... - Nature

Oncogene (2013) 32, 1121–1134 & 2013 Macmillan Publishers Limited All rights reserved 0950-9232/13 www.nature.com/onc

ORIGINAL ARTICLE

Phosphodiesterase-4 promotes proliferation and angiogenesis of lung cancer by crosstalk with HIF SS Pullamsetti1,2,5, GA Banat2,5, A Schmall1, M Szibor3, D Pomagruk2, J Ha¨nze4, E Kolosionek2, J Wilhelm2, T Braun3, F Grimminger2, W Seeger1,2, RT Schermuly1,2 and R Savai1 Lung cancer is the leading cause of cancer death worldwide. Recent data suggest that cyclic nucleotide phosphodiesterases (PDEs) are relevant in various cancer pathologies. Pathophysiological role of phosphodiesterase 4 (PDE4) with possible therapeutic prospects in lung cancer was investigated. We exposed 10 different lung cancer cell lines (adenocarcinoma, squamous and large cell carcinoma) to hypoxia and assessed expression and activity of PDE4 by real-time PCR, immunocytochemistry, western blotting and PDE activity assays. Expression and activity of distinct PDE4 isoforms (PDE4A and PDE4D) increased in response to hypoxia in eight of the studied cell lines. Furthermore, we analyzed various in silico predicted hypoxia-responsive elements (p-HREs) found in PDE4A and PDE4D genes. Performing mutation analysis of the p-HRE in luciferase reporter constructs, we identified four functional HRE sites in the PDE4A gene and two functional HRE sites in the PDE4D gene that mediated hypoxic induction of the reporter. Silencing of hypoxia-inducible factor subunits (HIF1a and HIF2a) by small interfering RNA reduced hypoxic induction of PDE4A and PDE4D. Vice versa, using a PDE4 inhibitor (PDE4i) as a cyclic adenosine monophosphate (cAMP) -elevating agent, cAMP analogs or protein kinase A (PKA)-modulating drugs and an exchange protein directly activated by cAMP (EPAC) activator, we demonstrated that PDE4-cAMP-PKA/EPAC axis enhanced HIF signaling as measured by HRE reporter gene assay, HIF and HIF target genes expression ((lactate dehydrogenase A), LDHA, (pyruvate dehydrogenase kinase 1) PDK1 and (vascular endothelial growth factor A) VEGFA). Notably, inhibition of PDE4 by PDE4i or silencing of PDE4A and PDE4D reduced human lung tumor cell proliferation and colony formation. On the other hand, overexpression of PDE4A or PDE4D increased human lung cancer proliferation. Moreover, PDE4i treatment reduced hypoxia-induced VEGF secretion in human cells. In vivo, PDE4i inhibited tumor xenograft growth in nude mice by attenuating proliferation and angiogenesis. Our findings suggest that PDE4 is expressed in lung cancer, crosstalks with HIF signaling and promotes lung cancer progression. Thus, PDE4 may represent a therapeutic target for lung cancer therapy. Oncogene (2013) 32, 1121–1134; doi:10.1038/onc.2012.136; published online 23 April 2012 Keywords: lung cancer; hypoxia; phosphodiesterases; proliferation; angiogenesis

INTRODUCTION Lung cancer is the leading cause of cancer-related death worldwide.1 The limited success of classic chemotherapeutics has led to focus on developing rationally targeted therapies aimed at the molecular mechanisms underlying tumorigenesis. Several novel drugs that specifically target these key components have been developed, but have yet to prove their efficacy for routine clinical application.2,3 Hence, it is essential to continue the identification of novel regulators of lung tumor growth and tumor microenvironment. Regulation of cyclic nucleotide signaling is regarded as a composite of multiple component pathways involved in diverse aspects of tumor cell function.4–7 The impairment of cyclic adenosine monophosphate (cAMP) or cyclic guanosine monophosphate (cGMP) generation by regulation of phosphodiesterases (PDEs) has been implicated in various cancer pathologies.8,9 PDEs are enzymes that regulate cellular levels of cAMP/cGMP by controlling their degradation. These enzymes are classified into 11 families (PDE1– PDE11) based on their sequence similarity, substrate preference and

sensitivity to various inhibitors.10,11phosphodiesterase 4 (PDE4), one of 11 PDE enzyme families, specifically catalyzes hydrolysis of cAMP. It has four subtypes (PDE4A, PDE4B, PDE4C and PDE4D) with at least 35 splice variants. PDE4 has a critical role in controlling intracellular cAMP concentrations.12 Importantly, many tumor cells originating from the central nervous system, lung and breast have PDE4 among all PDEs, as major regulator of cAMP-hydrolyzing activity.13 Interestingly, PDE4 inhibitors (PDE4i) were shown to inhibit growth and chemotaxis and to increase differentiation and/or apoptosis in pancreatic cancer cells, colon cancer cells and malignant melanoma cells.14–16 Inhibition of PDE4 by the selective inhibitor rolipram is shown to suppress tumor growth and augments the anti-tumor effects of chemotherapy and radiation therapy in medulloblastoma, glioblastoma17 and in several hematological malignancies.18 Why PDE4 isoforms are overexpressed in so many different tumor types remains elusive, but like other tumorigenesis factors, PDE4 isoforms may be regulated by changes in hypoxia. Adaptation to hypoxia is critical for tumor cell growth and

1 Department of Lung Development and Remodelling, Max-Planck-Institute for Heart and Lung Research, Bad Nauheim, Germany; 2Internal Medicine, University of Giessen Lung Center, Justus-Liebig-University, Giessen, Germany; 3Department of Cardiac Development and Remodelling, Max-Planck-Institute for Heart and Lung Research, Bad Nauheim, Germany; 4Department of Urology and Pediatric Urology, Philipps-University, Marburg, Germany and 5These authors contributed equally to this work. Correspondence: Dr R Savai, Molecular Mechanisms in Lung Cancer, Max Planck Institute for Heart and Lung Research, Parkstrasse-1, Bad Nauheim D-61231, Germany. E-mail: [email protected] Received 29 May 2011; revised 21 February 2012; accepted 24 February 2012; published online 23 April 2012

Phosphodiesterase 4 in lung cancer SS Pullamsetti et al

1122 survival.19 Hypoxia causes upregulation of genes through stabilization of hypoxia-inducible transcription factors (HIFs). HIF1a and HIF2a share common target genes characterized by a consensus hypoxia-responsive element (HRE), that are involved in cell growth, proliferation and angiogenesis.20 However, a direct relation between the hypoxia-HIF axis and the PDE4-cAMP axis in malignant cells has not been reported. We analyzed whether PDE4 expression is induced in hypoxia and involved in lung cancer progression. To evaluate this potential link, we studied the possible HIF dependence of PDE4. Furthermore, we analyzed the functional significance of predicted HREs (p-HREs) of the PDE4A and PDE4D genes. Also, we studied PDE4cAMP-mediated regulation of HIF and its downstream signaling molecules. In addition, we studied the influence of PDE4 inhibition on proliferation in vitro in a panel of 10 human lung cancer cells and assessed proliferation and angiogenesis in a nude mouse model of adenocarcinoma in vivo.

RESULTS PDE4 isoform expression increases in response to hypoxia To determine whether the expression of PDE4 isoforms (PDE4A, PDE4B, PDE4C and PDE4D) is induced in hypoxia, A549 adenocarcinoma cells were exposed to different oxygen concentrations (0.5, 1, 2 and 3% O2) for 12 and 24 h. Compared with PDE4 isoform expression in normoxia, the mRNA levels of both PDE4A and PDE4D increased significantly in response to hypoxia that is, with decreasing concentrations of O2 (Figure 1a). Real-time RT–PCR analysis revealed that PDE4A and PDE4D mRNA was markedly induced by hypoxic exposure for 12 and 24 h (Figure 1a). However, PDE4B is not expressed and PDE4C mRNA expression was decreased in response to hypoxia (Supplementary Figure 1a). Importantly, induction of PDE4A and PDE4D mRNA showed oxygen dependency. PDE4A and PDE4D mRNA was markedly induced by 0.5% O2, peaked by 1% O2 and decreased thereafter with increasing O2 concentrations (2 and 3%). Furthermore, immunoblotting confirmed that the protein levels of PDE4A (PDE4A1 74 kDa) and PDE4D (PDE4D4 105 kDa) were strongly induced by 24 h of exposure to 0.5 and 1% O2 concentrations (Figure 1b). As strong induction of both isoforms, PDE4A and PDE4D at mRNA and protein level occurred in A549 cells that were exposed to 1% O2, further experiments were continued with 1% O2 concentration. Immunofluorescence intensity measurements confirmed increased PDE4A- and PDE4D-like immunoreactivity in the cytoplasm after 24 h of exposure to hypoxia (Figures 1c–e). Interestingly, total cAMP-PDE activity was increased almost twofold in response to hypoxia compared with normoxic controls. Using PDE4i, we calculated the relative contribution of PDE4 activity to the total cAMP-PDE activity and found PDE4 to be the major contributor (Figure 1f). HIF1a or HIF2a is required for the regulation of PDE4A/PDE4D transcription in response to hypoxia Members of the HIF family of transcription factors are the major regulators of hypoxia-induced transcription. Hence, we studied whether hypoxia-mediated upregulation of PDE4 isoforms involves HIF. We examined expression and localization of HIF1a, HIF2a, PDE4A and PDE4D under normoxic and hypoxic conditions in A549 cells. Immunofluorescence analysis showed nuclear accumulation of HIF1a and HIF2a in hypoxia and an increase of PDE4A- and PDE4D-like immunoreactivity primarily in the cytoplasm (Figure 2a). Further, we studied whether PDE4A and PDE4D are transcriptional targets of HIF. To delineate, silencing of HIF1a and HIF2a by selective small interfering RNAs (siRNAs) was performed.21 In accordance, suppression of HIF1a and HIF2a protein expression Oncogene (2013) 1121 – 1134

in A549 cells transfected with selective siRNAs in hypoxia was confirmed by immunoblotting (Figure 2b). To determine whether hypoxic regulation of PDE4 involves HIF, we analyzed PDE4A or PDE4D mRNA expression in these transfected cells. Interestingly, we observed a significant downregulation of PDE4A in A549 cells transfected with either HIF1a or HIF2a siRNA, followed by exposure to hypoxia (Figure 2c). Similarly, PDE4D expression is suppressed by both HIF1a siRNA and HIF2a siRNA, compared with control siRNA (Figure 2c). Immunofluorescence as well as immunoblotting showed that PDE4A and PDE4D protein levels were significantly reduced after silencing HIF1a or HIF2a under hypoxic conditions (Figures 2d–f). In addition, selective silencing of HIF1a or HIF2a expression caused a significant reduction of total cAMP-PDE activity as well as PDE4 activity (Figure 2g), confirming the functional dependence of PDE4 on HIF1a. Hypoxia-HIF regulates PDE4A/PDE4D transcription in response to hypoxia To determine the molecular mechanisms underlying hypoxia- and HIF-induced PDE4A/PDE4D gene expression, we performed in silico analysis to identify rHRE located in the 50 regulatory regions of PDE4A and PDE4D (Figure 3a). We identified 13 p-HREs in PDE4A and 3 p-HREs in PDE4D. The p-HREs in PDE4A gene are located at 8210, 7133, 6970, 6851, 6387, 6159, 5460, 5016, 3439, 2425, 2063, 869 and 724, relative to the start codon site of PDE4A (Figure 3a). In PDE4D they are located at 7740, 4537 and 734 (Figure 3a). To verify that the p-HRE sites that are functional, we performed site-directed mutagenesis and generated a series of mutation constructs. To attain, we first assessed the hypoxic responsiveness of control plasmid (5 PGK-HRE HSV TKMP promoter) (Figure 3c). Further, 5 PGK-HRE’s in control vector (5 PGK-HRE HSV TKMP promoter) were replaced by the p-HRE regions (p-wtHRE; NCGTG±12N) or its mutated control sequences (p-mutHRE; NAAAG±12N) (Figure 3b). Transient transfection of p-wtHRE or p-mutHRE luciferase reporter constructs into A549 cells in presence of normoxia/ hypoxia demonstrated that only p-wtHRE constructs at 724, 869, 2063 and 5460 caused a significant induction of luciferase activity only in presence of hypoxia. Importantly, hypoxia was not able to stimulate luciferase activity of a construct where the p-HRE in the PDE4A/PDE4D promoter was mutated (p-mutHRE; Figure 3d). These results suggest that four HRE sites in PDE4A promoter are functional. On the other hand, two out of three predicted sites in PDE4D promoter were functional. p-wtHRE constructs at 4537 and 7740 showed 2–3 fold induction of luciferase activity in presence of hypoxia. However, no induction was observed when cells were transfected with the respective p-mutHRE (Figure 3e). PDE4-cAMP axis regulates HIF transcription and HIF target genes To investigate whether PDE4 regulates HIF-dependent signaling, A549 cells were transiently transfected with an HRE-dependent luciferase reporter plasmid and cultured under normoxic or hypoxic conditions in the absence or presence of PDE4i. Hypoxia caused a 20-fold induction of HRE-dependent luciferase activity that was reduced by addition of PDE4i (Figure 4a). Importantly, PDE4i also significantly reduced protein expression of HIF1a and HIF2a under hypoxic condition (Figure 4b). To delineate the downstream signaling involved in PDE4mediated HIF regulation, we examined the effect of forskolin (a cAMP-elevating agent) and 8-Br-cAMP (a cell-permeable, nonhydrolyzable cAMP analog) on HRE-dependent luciferase activity. Treatment with forskolin/8-Br-cAMP significantly decreased hypoxia-induced luciferase activity in A549 cells (Figures 4c and d). Having confirmed cAMP involvement, we further assessed cAMP-mediated signaling pathways, namely protein kinase A (PKA) and exchange protein directly activated by cAMP (EPAC), & 2013 Macmillan Publishers Limited


1123

Figure 1. PDE4 isoform mRNA and protein are overexpressed under hypoxia in A549 cells. A549 cells were cultured under normoxic or hypoxic conditions for 12 and 24 h. (a) Expression of PDE4 isoform mRNA (PDE4A and PDE4D) was assessed by real-time RT–PCR. The values were normalized to HPRT1 and relative changes were calculated using the DDCT method. (b) Protein expression was analyzed by immunoblotting with anti-PDE4A, anti-PDE4D and anti-b-actin and subsequently quantified. (c–e) Localization of PDE4A and PDE4D isoforms, assessed by measuring immunofluorescence intensity. Green (Alexa Fluor 488-conjugated PDE4A or PDE4D); blue (DAPI). Scale bars ¼ 100 mm. (f) Total cAMP-PDE activity in the absence or presence of PDE4i. All values represent the mean±s.e.m. (n ¼ 3–5). *Po0.05, **Po0.01 and ***Po0.001 versus normoxia.

using specific activators and inhibitors. Both PKA and EPAC activators (6-Benz-cAMP and 8-CPT-20 -O-Me-cAMP) decreased HRE-dependent luciferase activity under hypoxic conditions in A549 cells (Figures 4e and f). Interestingly, co-treatment with the PKA inhibitor KT5720 and PDE4i did not alter hypoxia-induced & 2013 Macmillan Publishers Limited

HRE-dependent luciferase activity (Supplementary Figure 2). Moreover, PDE4i significantly decreased the expression of HIFdependent target genes. Hypoxia-induced mRNA expression of lactate dehydrogenase A (LDHA), pyruvate dehydrogenase kinase 1 (PDK1) and vascular endothelial growth factor (VEGF) decreased Oncogene (2013) 1121 – 1134


1124

Oncogene (2013) 1121 – 1134

& 2013 Macmillan Publishers Limited


1125 after PDE4i treatment (Figure 4g). In line, VEGF protein levels measured by immunoblotting demonstrated dose-dependent decrease of VEGF expression after PDE4i treatment (Figure 4h). PDE4 inhibition reduces lung cancer cell proliferation and VEGF secretion As PDE4i regulates VEGF expression, we determined the effects of PDE4i on VEGF secretion. As expected, VEGF concentration in culture supernatants, which was enhanced under hypoxic conditions, was significantly reduced by PDE4 inhibition (Supplementary Figure 3). To confirm the role of PDE4 in proliferation of human lung tumor cells, A549 cells in normoxia or hypoxia were treated with PDE4i. Cell proliferation was quantified by incorporation of [3H]thymidine or BrdU. A549 cells exposed to hypoxia for 24 h incorporated significantly more [3H]thymidine and BrdU compared with cells incubated under normoxic conditions (data not shown), indicating that hypoxiainduced tumor cell proliferation. Addition of PDE4i to A549 cells reduced hypoxia-stimulated [3H]thymidine and BrdU incorporation in a concentration-dependent manner. However, the antiproliferative effects of PDE4i were less pronounced in A549 cells exposed to normoxia (Figures 5a and b). Given this marked effect of PDE4i on cell proliferation under hypoxic conditions, we then investigated the effect of PDE4 inhibition on the expression of key control elements of the G1/Sphase transition, including cyclin B1, cyclin E and CDK inhibitor p27, in A549 cells treated with PDE4i. As shown in Supplementary Figure 4, PDE4i treatment decreased the expression of cyclins B1 and E compared with control cells. By contrast, p27 expression level did not change significantly (Supplementary Figure 4). The growth inhibition of PDE4i on lung cancer cells was also assessed using colony formation assays. Importantly, the results were similar to those obtained from the proliferation assays. Compared with the control, PDE4i decreased the number of colonies in a concentration-dependent manner. PDE4i decreased the colony numbers by 75% (10 mM) and 57% (50 mM), respectively (Figures 5c and d).On the other hand, PDE4i strongly decreased the colony numbers in A549 cells exposed to hypoxia that is, by 67% (10 mM) and 45% (50 mM), respectively (Figures 5c and d). PDE4 inhibits proliferation and HIF signaling of several human lung cancer cell lines We next investigated whether PDE4 activation, the anti-proliferative effects of PDE4i as well as the effects of PDE4i on HIF signaling are restricted to A549 cells or have a broader impact on other human non-small-cell lung cancer cells. Investigation of the anti-proliferative effects of PDE4i in nine different human lung cancer cells (five adenocarcinomas, two large cell carcinomas and two squamous cell carcinomas) exposed to normoxia or hypoxia revealed that five of them were highly sensitive to PDE4i (H1299, H1975, H661, A520 and H1437), two of them displayed intermediate sensitivity (H23 and H1650), and two of them were resistant to PDE4i (H460 and H226) (Figures 6a and 7a). We measured total cAMP-hydrolyzing PDE activity and PDE4 activity in all the nine different human lung cancer cells. In seven of the nine cell lines, PDE4 activity contributed significantly to the total cAMP-PDE activity (Figures 6b and 7b).

Furthermore, we assessed the influence of PDE4i on protein expression of HIF isoforms and HIF downstream target, VEGF. We observed that HIF1a, HIF2a and VEGF were down regulated by PDE4i in H1299, H661, H1437 and H1975 cells in dose-dependent manner (Figures 6c and 7c). PDE4i downregulated HIF and VEGF only at a high concentration in H23, H1650 and A520 cells (Figures 6c and 7c). PDE4A- or PDE4D-selective siRNA reduces lung cancer cell proliferation To more specifically evaluate the PDE4 inhibition, pooled siRNAs targeting PDE4A and PDE4D were used, that were shown to selectively knock down PDE4A and PDE4D, respectively.22 Each of PDE4A- and PDE4D-siRNA knocked down the expression of its corresponding endogenous mRNA in hypoxia (Figure 8a). Downregulation of PDE4A and PDE4D protein expression was also confirmed by immunoblotting (Figure 8b). Suppression of PDE4A and PDE4D protein expression was also confirmed by immunocytochemistry (Supplementary Figure 5). Further, to determine the effects on proliferation, PDE4A or PDE4D siRNA was transfected into A549 cells followed by exposure to normoxia or hypoxia for 24 h. Transfection of A549 cells with either PDE4A siRNA or PDE4D siRNA caused a significant decrease in hypoxia-stimulated [3H]thymidine and BrdU incorporation. Control siRNA had no effect on proliferation (Figures 8c and d). Interestingly, transfection of both PDE4A and PDE4D siRNAs showed additive effects on proliferation suppression compared with PDE4A or PDE4D siRNAs alone (Figure 8b–d). PDE4A- or PDE4D overexpression induces lung cancer cell proliferation To unveil that whether PDE4 isoforms itself induces proliferation in A549 cells, A549 cells were transiently transfected with PDE4A, PDE4D or both PDE4A and PDE4D expression plasmid or empty vector. Real-time RT–PCR indicated that PDE4A or PDE4D mRNA was dramatically increased in PDE4 constructs transfected cells compared with empty vector transfected cells (Figure 8e). Moreover, we observed an overexpression of PDE4A or PDE4D protein in these cells (Figure 8f). Further assessing whether PDE4 was able to induce proliferation in presence of normoxia/hypoxia in A549 cells, we observed that PDE4A-, PDE4D- overexpression or combination of both caused an increase in both normoxia and hypoxia-stimulated [3H]thymidine and BrdU incorporation (Figures 8g and h). Under hypoxic conditions the proliferation induced by PDE4A or PDE4D overexpression is much stronger compared with respective normoxic controls (Figures 8g and h). PDE4i inhibits A549 tumor xenograft growth in nude mice by altering proliferation and angiogenesis To study the effects of PDE4 inhibition in vivo, A549 cells were injected s.c. into nude mice. When tumor size reached B36 mm3, mice were treated with PDE4i at different doses (1, 3 or 6 mg/kg) i.p. every day until day 36. We did not observe any significant change in body weight (Supplementary Figure 6) or any adverse behavioral effects in animals treated with PDE4i compared with the control animals throughout the experiment. Regarding its anticancer efficacy, PDE4i treatment at 3 and 6 mg/kg started

Figure 2. PDE4 isoform overexpression under hypoxia in A549 cells is dependent on HIF. (a) Localization of HIF1a, HIF2a, PDE4A and PDE4D was assessed by immunofluorescence in A549 cells that were cultured under normoxia or hypoxia for 24 h. Confocal microscopic images are shown. (b–h) A549 cells were transiently transfected with HIF1a siRNA, HIF2a siRNA or scrambled siRNA (si-control). siRNA-transfected A549 cells were cultured under normoxia or hypoxia. (b) After 24 h, protein expression of HIF1a and HIF2a by immunoblotting. (c) mRNA expression of PDE4A and PDE4D was assessed by real-time RT–PCR. (d) Protein expression of PDE4A and PDE4D was assessed by immunoblotting. (e) Immunofluorescence and (f, g) the intensities of the fluorescent signals were determined by Leica DMLA and QWin 500IW software (Leica, Wetzlar, Germany). Green (Alexa Fluor 488-conjugated PDE4A or PDE4D); blue (DAPI). Scale bars ¼ 100 mm. (h) Total cAMP-PDE activity was measured. All values represent the mean±s.e.m. (n ¼ 3–5). ***Po0.001 versus si-control. & 2013 Macmillan Publishers Limited

Oncogene (2013) 1121 – 1134


1126

Figure 3. Identification of HREs in PDE4A and PDE4D that mediate gene induction in hypoxia. (a) Schematic representation of the hypoxic response element (HRE) binding sites within the PDE4A (blue) or PDE4D (red) promoter region. The predicted HREs are shown with arrow. Putative consensus-binding sites for selected transcription factor family matrices were identified using the HUSAR Bioinformatics Laboratory software. Sequences for putative binding sites and their positions are indicated. (b) Schematic representation of wild-type (p-wtHRE) or mutated (p-mutHRE) HSV TKMP promoter constructs generated. (c–e) A549 cells were transfected with either (c) control plasmid (5 PGKHRE HSV TKMP promoter), (d) p-wtHRE and p-mutHRE containing PDE4A promoter reporter construct or (e) p-wtHRE and p-mutHRE containing PDE4D promoter reporter construct. 6 h after transfection, A549 cells were cultured under normoxia or hypoxia for 24 h. Relative luciferase activity in cell extracts was measured with a luminometer and is expressed as relative light units. All values represent the mean±s.e.m. (n ¼ 6). *Po0.05 versus p-wtHRE/normoxia; **Po0.01 versus p-wtHRE/normoxia; ***Po0.001 versus p-wtHRE/normoxia; yy Po0.01 versus p-wtHRE/hypoxia; yyyPo0.001 versus p-wtHRE/hypoxia.

showing inhibition of tumor growth at the 8th day of treatment, which became more visible and statistically significant at 24th day (Figure 7a). At the study end (36th day), tumor size decreased from 1909.02±79.28 mm3 (94.25%) in placebo group to Oncogene (2013) 1121 – 1134

888.61±196.72 mm3 (43.87%) in 1 mg/kg, 831.33±144.49 mm3 (41.04%) in 3 mg/kg and 706.68±136.86 mm3 (34.89%) in 6 mg/kg PDE4i-treated groups, respectively (Figure 9a). The subcutaneous tumor size is shown in representative photographs in Figure 9b & 2013 Macmillan Publishers Limited


1127

Figure 4. PDE4i and cAMP agonists inhibits HIF. A549 cells were transfected with HRE reporter plasmids and treated with (a) PDE4i (rolipram; 1, 10, and 50 mM), (c) forskolin (50, 100 and 500 mM), (d) 8-Br-cAMP (100, 500 and 1 mM), (e) 8-CPT-20 -O-Me-cAMP (50, 100 and 200 mM) or (f) 6-BenzcAMP (50, 100 and 200 mM). At 6 h after transfection, A549 cells were cultured under normoxia or hypoxia for 24 h. Relative luciferase activity in cell extracts was measured with a luminometer and is expressed as relative light units. All values represent the mean±s.e.m. (n ¼ 6). ***Po0.001 versus control. (b) Protein expression of HIF1a, HIF2a and -b-actin of A549 cells that were exposed to hypoxia and treated with PDE4ifor 24 h (rolipram; 1, 10 and 50 mM). (g) mRNA expression of HIF-dependent genes (LDHA, PDK1 and VEGF) was assessed by real-time RT–PCR, values were normalized to HPRT1 and relative changes were calculated using the DDCT. *Po0.05 versus control; **Po0.01 versus control; ***Po0.001. (h) Protein expression of VEGF and -b-actin of A549 cells that were exposed to hypoxia and treated with PDE4i for 24 h (rolipram; 1, 10 and 50 mM).

(left). To further study the significance of our cell culture findings related to growth inhibition by PDE4i in vivo, we performed Ki67 staining in PDE4i-treated tumor tissue. The number of proliferating cells in tumors stained with Ki67 was reduced in PDE4i-treated mice (61.59% in 1 mg/kg; 49.65% in 3 mg/kg and 40.60% in 6 mg/ kg) compared with control mice or mice treated with a placebo (Figures 9b and c). In addition, analysis of angiogenesis in tumors by staining vessels for CD31 indicated decreased angiogenesis in PDE4i-treated mice compared with control mice or mice treated with a placebo (Figures 9b and d). & 2013 Macmillan Publishers Limited

DISCUSSION We report hypoxia-induced regulation of PDE4 in most lung tumor cell types, including adenocarcinoma, large cell carcinoma and squamous cell carcinoma. Upregulation of PDE4 isoforms (mainly PDE4A and PDE4D) by hypoxia was mediated by HIF1a and HIF2a. Furthermore, PDE4 regulated HIF transcriptional activity via a cAMP-PKA/EPAC-dependent pathway. PDE4 inhibition by rolipram or siRNA-mediated knockdown of PDE4 reduced tumor cell proliferation and altered the expression of cell-cycle genes. PDE4i potently reduced lung tumor growth in A549 lung tumor Oncogene (2013) 1121 – 1134


1128

Figure 5. PDE4 inhibition reduces proliferation of A549 cells under hypoxia. Serum-starved A549 cells were pretreated with PDE4i (rolipram; 1, 10 or 50 mM) and exposed to normoxia or hypoxia for 24 h. (a) Cell proliferation was assessed by [3H]thymidine incorporation. (b) Cell proliferation was assessed by BrdU incorporation. (c, d) Colony-forming ability of A549 cells in the groups treated as described above was analyzed by the colony formation assay under normoxia. (c, d) Representative plates and the manual quantification of colonies made from five independent plates. All values represent the mean±s.e.m. (n ¼ 4–5). *Po0.05 and ***Po0.001 versus vehicle-treated group (control).

xenografts in athymic nude mice by altering proliferation and angiogenesis (Figure 9e). PDE4 has been reported as a major cAMP-hydrolyzing enzyme in several tumor cell lines.13 In the current study, we confirmed the leading role of PDE4 in cAMP hydrolysis in several human nonsmall-cell lung cancer cell lines. As it is well known that hypoxia has an important role in lung tumor progression by regulating HIF1 transcription factor.23,24 We investigated whether PDE4 isoforms represent a direct target of hypoxia. Indeed, we noted increased mRNA and protein expression of PDE4A and PDE4D isoforms in response to hypoxia, which was confirmed by increased PDE4 activity. Several predicted HREs were identified in the 50 regulatory regions of PDE4A and PDE4D.25 Importantly, generating several wild-type/mutated HRE luciferase reporter constructs, we identified four functional HREs in the PDE4A 50 regulatory gene region and two functional HREs in PDE4D 50 regulatory gene region. Moreover, we demonstrated that the observed increase in PDE4A and PDE4D levels in hypoxia was dependent on HIF1a and HIF2a, as siRNA-mediated silencing of HIF1a and HIF2a led to decreased expression and activity of PDE4A and PDE4D isoforms. To our knowledge, only two studies have reported a potential hypoxia-induced activation of PDE4 in nontumorigenic cells.26,27 Millan et al.26 showed a transient upregulation of PDE4B and a sustained upregulation of PDE4A and PDE4D upon exposure of human pulmonary artery smooth muscle cells to hypoxia.26 Also, Nunes et al.27 detected enhanced PDE4 activity in the rat carotid body induced by acute hypoxia. Yet, ours is the first study to explicitly address PDE4 induction by hypoxia in lung tumor cells and assess the underlying molecular mechanisms. Oncogene (2013) 1121 – 1134

Intriguingly, we found that PDE4 inhibition decreased HIF transcriptional activity in A549 cells only under hypoxic conditions. This finding was strengthened by decreased protein expression of HIF and mimicked by results obtained from treating hypoxic A549 cells with a cAMP-elevating agent (forskolin), cAMP analog (8-BrcAMP), PKA activator (6-Benz-cAMP) or EPAC activator (8-pCPT-20 O-Me-cAMP) agents that also decreased HIF transcriptional activity. In addition, PDE4 inhibition suppressed induction of hypoxia-responsive genes such as LDHA, PDK1 and VEGFA. Interestingly, PDE4i led to decrease in the amount of secreted VEGF compared with A549 cells that were exposed to hypoxia. Hence, our study demonstrates pronounced crosstalk between HIF—a master regulator of hypoxic responses—and the PDE4 – cAMP pathway. As PDE4 isoforms are overexpressed under hypoxic conditions and could be implicated in lung tumor progression, we first assessed the role of PDE4 isoforms in proliferation of lung cancer cells. We demonstrated that PDE4i or PDE4A/PDE4D silencing significantly decreased the proliferation and colony-forming abilities of lung cancer cells under hypoxia and normoxia, but the effects were more pronounced under hypoxia. In addition, we observed decreased expression of cyclin B1 and cyclin E in the presence of PDE4i. Most importantly, the anti-proliferative effects of PDE4 inhibition were observed in a panel of human lung cancer cells that included adenocarcinomas, large cell carcinomas and squamous cell carcinomas. This effect is intriguing and suggests the widespread therapeutic potential of PDE4i. These data are in agreement with previous studies that demonstrated anti-proliferative effects of PDE4 inhibition in human glioma, acute lymphoblastic leukemia and pancreatic cancer cells.14,28,29 & 2013 Macmillan Publishers Limited


1129

Figure 6. PDE4 inhibition reduces proliferation in a panel of human lung cancer cells. Serum-starved human lung cancer cells (H1299, H661, H460, H226 and H1437) were pretreated with the PDE4i and exposed to normoxia or hypoxia for 24 h. (a) Cell proliferation was assessed by [3H]thymidine incorporation assay. (b) Total cAMP-PDE activity was measured in hypoxia or in normoxia in the presence or absence of PDE4i. (c) Protein expression of HIF1a, HIF2a and b-actin. All values represent the mean±s.e.m. (n ¼ 3–5). *Po0.05, **Po0.01, and ***Po0.001 versus vehicle-treated group (control).


Oncogene (2013) 1121 – 1134


1130

Figure 7. PDE4 inhibition reduces proliferation in a panel of human lung cancer cells. Serum-starved human lung cancer cells (H1975, H520, H23 and H1650) were pretreated with the PDE4i and exposed to normoxia or hypoxia for 24 h. (a) Cell proliferation was assessed by [3H]thymidine incorporation assay. (b) Total cAMP-PDE activity was measured in hypoxia or in normoxia in the presence or absence of PDE4i. (c) Protein expression of HIF1a, HIF2a, VEGF and b-actin. All values represent the mean±s.e.m. (n ¼ 3–5). *Po0.05, **Po0.01, and ***Po0.001 versus vehicle-treated group (control).

In fact, the growth inhibitory effects of PDE4i appear to be celltype specific; in the melanoma cell line HMG, the PDE4i rolipram and Ro-20-1724 promote cell growth.15 Consistent with these in vitro studies, PDE4i treatment showed significant anti-tumor efficacy on A549 human lung cancer xenografts in nude mice. To our knowledge, this is the first report to show a potential therapeutic role for PDE4 inhibition in the treatment of lung cancer. PDE4 is currently being developed as a novel therapeutic target for the treatment of asthma and chronic obstructive pulmonary disease.30,31 The anti-tumor effects of PDE4i are most readily attributed to decreased cell proliferation, as evidenced by decreased Ki67-positive cells. This can also be attributed to the anti-angiogenic potential of PDE4i as shown by Oncogene (2013) 1121 – 1134

the reduced number of CD31-positive vessels in lung tumors. This is of potential importance because angiogenesis in many solid tumors, including lung cancer, is a major factor in predicting disease aggressiveness and patient survival.32 In addition, we have previously demonstrated that PDE4i attenuates the epithelialmesenchymal transition in A549 cells,22 which has been hypothesized to be a critical event in the invasion and metastasis of tumor cells. In summary, we report that PDE4 is a positive regulator of the hypoxia-HIF signaling pathway and has an important role in proliferation of several human lung cancer cell lines. PDE4 inhibition suppressed lung tumor growth in vivo by decreasing cell proliferation and angiogenesis. These findings suggest PDE4 as a novel & 2013 Macmillan Publishers Limited


1131

Figure 8. siRNA-mediated knockdown or overexpression of PDE4A and PDE4D regulates proliferation of A549 cells under hypoxia. (a–d) A549 cells were transiently transfected with PDE4A siRNA, PDE4D siRNA, combination PDE4A þ PDE4D siRNA or scrambled siRNA (si-control). siRNAtransfected cells were cultured under normoxia or hypoxia for 24 h. (a) Expression of PDE4 isoform mRNA (PDE4A and PDE4D) was assessed by real-time RT–PCR, values was normalized to HPRT1, and relative changes were calculated using the DDCT method. (b) Protein expression of PDE4A, PDE4D and b-actin as assessed by immunoblotting. (c) Cell proliferation was assessed by [3H]thymidine incorporation. (d) Cell proliferation was assessed by BrdU incorporation. (e–h) A549 cells were transiently transfected with PDE4A, PDE4D or combination PDE4A þ PDE4D expression plasmids or empty vector. Transfected cells were cultured under normoxia or hypoxia for 24 h. (e) Expression of PDE4 isoform mRNA (PDE4A and PDE4D) was assessed by real-time RT–PCR, values was normalized to HPRT1, and relative changes were calculated using the DDCT method. (f) Protein expression of PDE4A, PDE4D and b-actin as assessed by immunoblotting. (g) Cell proliferation was assessed by [3H]thymidine incorporation. (h) Cell proliferation was assessed by BrdU incorporation. All values represent the mean±s.e.m. (n ¼ 3–5). ***Po0.001 versus si-control.


Oncogene (2013) 1121 – 1134


1132

Figure 9. PDE4 inhibition decreases tumor growth in A549 adenocarcinoma mouse model. A549 tumor-bearing Balb/c nude mice were treated with an i.p. injection of placebo or PDE4i (1, 3 and 6 mg/kg) and control. (a) Tumor size (mm3) was measured over time after PDE4i treatment shown. Results were expressed as % of control, n ¼ 12, Po0.05, Po0.01 and Po0.001 versus placebo. (b, left) Representative photographs of mice with subcutaneous tumors in the different treatment groups. (b, middle and right) Representative immunohistochemical microphotographs of Ki67 þ and CD31 þ stained tumor sections. Scale bars ¼ 20 mm. (c) Quantification of Ki67 þ nuclei, expressed as % of control. (d) Quantification of CD31 þ vessels, expressed as % of control. All values represent the mean±s.e.m. (n ¼ 8) **Po0.01 and ***Po0.001 versus placebo. (e) Diagram of signaling events leading to reduced tumor growth by PDE4 inhibition. PDE4 inhibition with the use of either a chemical inhibitor (rolipram) or gene silencing (PDE4A/D siRNA) increases intracellular cAMP levels. Increased cAMP decreases HIF transcriptional activity through PKA and EPAC. This in turn reduces the expression of the HIF target genes LDHA, PDK1 and VEGFA as well as the expression of cyclins. Collectively, all these changes lead to reduced proliferation and angiogenesis and ultimately to a reduction in lung tumor growth.

molecular target for lung cancer therapy and indicate that PDE4 inhibition should be evaluated in clinical trials for lung tumors.

and H520) cell lines were obtained from American Type Culture Collection (Table 1) (Manassas, VA, USA). See data supplement for culture conditions.

Hypoxia exposure and treatment MATERIALS AND METHODS Cell lines Human adenocarcinoma (A549, H1299, H1437, H1650, H1975 and H23), large cell carcinoma (H460 and H661) and squamous cell carcinoma (H226 Oncogene (2013) 1121 – 1134

A549 adenocarcinoma cells were exposed to different oxygen concentrations (0.5, 1, 2 and 3% O2) for 12 and 24 h and studied the regulation of PDE4 isoforms. All other experiments were performed with 1% O2 or normoxia. For proliferation, PDE activity, or reporter gene assays, A549 cells were treated with (i) a selective PDE4i(rolipram), (ii) the cAMP-elevating & 2013 Macmillan Publishers Limited


1133 Table 1.

Cancer cell lines used in this study

Sample Name

Provider

Histology

A549 H1299 H1437 H1650

ATCC ATCC ATCC ATCC

H1975 H23 H460 H520 H661 H226

ATCC ATCC ATCC ATCC ATCC ATCC

Carcinoma Carcinoma, not specified Adenocarcinoma Adenocarcinoma, bronchoalveolar carcinoma Adenocarcinoma Adenocarcinoma Large cell lung cancer Squamous cell carcinoma Large cell lung cancer Squamous cell carcinoma; mesothelioma

TNM Stage

Metastatic site of derivation

I IIIB

Lymph node Pleural effusion Pleural effusion

Isolation date

Growth (adh/susp)

Morphology

1972

Adherent Adherent Adherent Adherent

Epithelial Epithelial

June, 1986 May, 1987 July, 1988

Pleural effusion Lymph node Pleural effusion

agent, forskolin, (iii) the cAMP analog, 8-Br-cAMP, (iv) the PKA activator, 6-Benz-cAMP, (v) the PKA inhibitor, KT5720, and (vi) the EPAC activator, 8-CPT-20 -O-Me-cAMP. For vehicle (control), we used DMSO or dH2O, respectively. See online data supplement for details.

RNA Interference We used siRNA oligonucleotides specific for PDE4A, PDE4D, HIF1a, HIF2a and scramble siRNA (si-control). Sequences of siRNA are provided in data supplement. Transient transfection with siRNAs was performed with Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s protocols. See online data supplement for details.

1982 1982 1982 March, 1980

Adherent Adherent Adherent Adherent Adherent Adherent

Gender

Age

Epithelial

M M M M

58 43 60 27

Epithelial Epithelial Epithelial Epithelial Epithelial

F M M M M M

51 43

Colony formation assays Two hundred cells were seeded in a 60-mm culture dish followed 24 h later by incubation without or with PDE4i. After incubation for 10 days in normoxia and hypoxia, the colonies were stained with crystal violet, rinsed in water, airdried and photographed. Manual counting scored colonies containing more than 50 cells, and five plates were counted per treatment group.

Vascular endothelial growth factor (VEGF) measurement A549 cells were exposed to normoxia or hypoxia in the absence or presence of PDE4i. Cell supernatants were collected, clarified by centrifugation at 2000 r.p.m. for 5 min, and measured for VEGF concentration in the supernatant using an ELISA kit (R&D Systems, Minneapolis, MN, USA).

RNA isolation and real-time RT–PCR Total RNA was extracted from the hypoxia-exposed and siRNA-treated cells with Trizol Reagent. Total RNA was reverse transcribed with ImProm-II Reverse Transcription System, followed by RT–PCR analysis of PDE4A, PDE4B, PDE4C, PDE4D, HPRT1, LDHA, PDK1 and VEGF. For details of protocol and primer sequences, see data supplement.

Animals Immunodeficient female BALB/c nu/nu mice 7–8 weeks old were purchased from Charles Rivers (Sulzfeld, Germany), kept under pathogen-free conditions, and handled in accordance with the European Communities recommendations for experimentation.

PDE activity assays PDE activity was determined with a radio enzymatic assay as described in data supplement. Total cAMP-PDE activity was assessed at 1 mM cAMP and the contribution of PDE4 isozymes was determined by using selective inhibitor-rolipram.

Immunoblotting Immunoblotting was performed for detection of b-actin, HIF1a, HIF2a, PDE4A, PDE4D, VEGF, p27, cyclin E and cyclin B1 from protein lysates. See data supplement for details.

Tumorigenicity in nude mice Tumorigenicity was assessed by subcutaneous injection of A549 cells (5 106 cells/200 ml in saline) into female BALB/c nu/nu mice. Treatment was initiated when tumors were 35.95±3.71 mm3. The selective PDE4i rolipram (1, 3 and 6 mg/kg, with vehicle (DMSO) as a control) was administered s.c. once daily for 36 days. Tumor size was calculated using the equation: length (width)2)/2 as described.34 After 36 days of treatment, the mice were anesthetized for imaging and killed.

Immunostaining Construction of HRE plasmids and luciferase assay The pmirGLO vector (Promega, Mannheim, Germany) was modified to test HRE sites that were predicted in silico within the 8 kb 50 upstream region of PDE4A (ENST00000344979) and PDE4D (ENST00000360047) genes. Initially, the polylinker was replaced by the fivefold tandem repeat of the PGK-HRE coupled to the HSV-TKmp promoter as described and analyzed.33 The resulting plasmid served as a control construct and as origin for insertion of the predicted PDE4A and PDE4D HRE regions (NCGTG±12N) or its mutated control sequences (NAAAG±12N). A549 cells were transfected by Lipofectamine 2000. The dualluciferase analysis of promoter activities was performed according to the manufacturer‘s instructions (Promega) employing a Spectrofluorometer (BioTek Instruments GmbH, Bad Friedrichshall, Germany).

Proliferation assays Proliferation assays were performed with tumor cells exposed to different experimental conditions: (i) tumor cells transfected with siPDE4A, siPDE4Dor both siPDE4A/siPDE4D, (ii) tumor cells transfected with plasmid construct encoding PDE4A, PDE4D or both PDE4A/PDE4D and (ii) tumor cells treated with PDE4i. After 20 h treatment, [3H]thymidine or BrdU was added. Thymidine or BrdU incorporation was determined after 4 h by beta counter or ELISA reader, respectively, as described in data supplement. & 2013 Macmillan Publishers Limited

A549 cells grown on chamber slides were treated and fixed with acetonemethanol (1:1). After blocking with 5% BSA, cells were incubated with the following primary antibodies, PDE4A, PDE4D, HIF1a and HIF2a. Indirect immunofluorescence was conducted by incubation with Alexa 488-conjugated secondary antibodies (Molecular Probes Inc., Eugene, OR, USA). For immunohistochemical staining, consecutive 5 mm cryostat s.c. tumor sections were fixed with acetone-methanol and then hydrated in PBS. Tissues were blocked in 5% BSA/PBS and incubated with primary antibody against proliferating cell nuclear antigen (Ki67) or CD31. For indirect immunofluorescence, slides were incubated with Alexa Fluor 488-labeled secondary antibodies. After incubation, all sections were counterstained with DAPI (for nuclear staining) and mounted with fluorescent mounting media (Dako Cytomation, Glostrup, Denmark). At the end of the procedure, fluorescence images were acquired. Details of imaging parameters and the fluorescent intensity quantified as described in data supplement.

Data analysis All graphs and statistical procedures were done using GraphPad 5 (San Diego, CA, USA). The analysis of variance one-way test followed by the Dunnett post hoc was used for data analysis. Data are expressed as mean±s.e. Statistical significance was set at Po0.05. Oncogene (2013) 1121 – 1134


1134 CONFLICT OF INTEREST The authors declare that they have no conflict of interest.

ACKNOWLEDGEMENTS The authors thank Marianne Hoeck and Katharina Weidl for her excellent technical assistance. Dr Rajkumar Savai’s research was supported by the STARTUP AWARD from Justus-Liebig University Giessen, Germany and LOEWE-Center UGMLC (Universities of Giessen and Marburg Lung Center).

REFERENCES 1 Jemal A, Siegel R, Ward E, Hao Y, Xu J, Thun MJ. Cancer statistics 2009CA Cancer J Clin 2009; 59: 225–249. 2 Pallis AG, Serfass L, Dziadziusko R, van Meerbeeck JP, Fennell D, Lacombe D et al. Targeted therapies in the treatment of advanced/metastatic NSCLC. Eur J Cancer 2009; 45: 2473–2487. 3 Rossi A, Maione P, Colantuoni G, Ferrara C, Rossi E, Guerriero C et al. Recent developments of targeted therapies in the treatment of non-small cell lung cancer. Curr Drug Discov Technol 2009; 6: 91–102. 4 Yamanaka Y, Mammoto T, Kirita T, Mukai M, Mashimo T, Sugimura M et al. Epinephrine inhibits invasion of oral squamous carcinoma cells by modulating intracellular cAMP. Cancer Lett 2002; 176: 143–148. 5 Timoshenko AV, Xu G, Chakrabarti S, Lala PK, Chakraborty C. Role of prostaglandin E2 receptors in migration of murine and human breast cancer cells. Exp Cell Res 2003; 289: 265–274. 6 McEwan DG, Brunton VG, Baillie GS, Leslie NR, Houslay MD, Frame MC. Chemoresistant KM12C colon cancer cells are addicted to low cyclic AMP levels in a phosphodiesterase 4-regulated compartment via effects on phosphoinositide 3-kinase. Cancer Res 2007; 67: 5248–5257. 7 Deguchi A, Thompson WJ, Weinstein IB. Activation of protein kinase G is sufficient to induce apoptosis and inhibit cell migration in colon cancer cells. Cancer Res 2004; 64: 3966–3973. 8 Zhang L, Murray F, Zahno A, Kanter JR, Chou D, Suda R et al. Cyclic nucleotide phosphodiesterase profiling reveals increased expression of phosphodiesterase 7B in chronic lymphocytic leukemia. Proc Natl Acad Sci USA. 2008; 105: 19532–19537. 9 Marko D, Romanakis K, Zankl H, Furstenberger G, Steinbauer B, Eisenbrand G. Induction of apoptosis by an inhibitor of cAMP-specific PDE in malignant murine carcinoma cells overexpressing PDE activity in comparison to their nonmalignant counterparts. Cell Biochem Biophys 1998; 28: 75–101. 10 Omori K, Kotera J. Overview of PDEs and their regulation. Circ Res 2007; 100: 309–327. 11 Conti M, Beavo J. Biochemistry and physiology of cyclic nucleotide phosphodiesterases: essential components in cyclic nucleotide signaling. Annu Rev Biochem 2007; 76: 481–511. 12 Conti M, Richter W, Mehats C, Livera G, Park JY, Jin C. Cyclic AMP-specific PDE4 phosphodiesterases as critical components of cyclic AMP signaling. J Biol Chem 2003; 278: 5493–5496. 13 Marko D, Pahlke G, Merz KH, Eisenbrand G. Cyclic 30 ,50 -nucleotide phosphodiesterases: potential targets for anticancer therapy. Chem Res Toxicol 2000; 13: 944–948. 14 Mouratidis PX, Colston KW, Bartlett JB, Muller GW, Man HW, Stirling D et al. Antiproliferative effects of CC-8062 and CC-8075 in pancreatic cancer cells. Pancreas 2009; 38: 78–84. 15 Narita M, Murata T, Shimizu K, Nakagawa T, Sugiyama T, Inui M et al. A role for cyclic nucleotide phosphodiesterase 4 in regulation of the growth of human malignant melanoma cells. Oncol Rep 2007; 17: 1133–1139.

16 Murata K, Sudo T, Kameyama M, Fukuoka H, Muka M, Doki Y et al. Cyclic AMP specific phosphodiesterase activity and colon cancer cell motility. Clin Exp Metastasis 2000; 18: 599–604. 17 Goldhoff P, Warrington NM, Limbrick Jr DD, Hope A, Woerner BM, Jackson E et al. Targeted inhibition of cyclic AMP phosphodiesterase-4 promotes brain tumor regression. Clin Cancer Res 2008; 14: 7717–7725. 18 Lerner A, Epstein PM. Cyclic nucleotide phosphodiesterases as targets for treatment of haematological malignancies. Biochem J 2006; 393(Pt 1): 21–41. 19 Harris AL. Hypoxia–a key regulatory factor in tumour growth. Nat Rev Cancer 2002; 2: 38–47. 20 Semenza GL. Regulation of physiological responses to continuous and intermittent hypoxia by hypoxia-inducible factor 1. Exp Physiol 2006; 91: 803–806. 21 Sowter HM, Raval RR, Moore JW, Ratcliffe PJ, Harris AL. Predominant role of hypoxia-inducible transcription factor (Hif)-1alpha versus Hif-2alpha in regulation of the transcriptional response to hypoxia. Cancer Res 2003; 63: 6130–6134. 22 Kolosionek E, Savai R, Ghofrani HA, Weissmann N, Guenther A, Grimminger F et al. Expression and activity of phosphodiesterase isoforms during epithelial mesenchymal transition: the role of phosphodiesterase 4. Mol Biol Cell 2009; 20: 4751–4765. 23 Swinson DE, O’Byrne KJ. Interactions between hypoxia and epidermal growth factor receptor in non-small-cell lung cancer. Clin Lung Cancer 2006; 7: 250–256. 24 Hung JJ, Yang MH, Hsu HS, Hsu WH, Liu JS, Wu KJ. Prognostic significance of hypoxia-inducible factor-1alpha, TWIST1 and Snail expression in resectable nonsmall cell lung cancer. Thorax 2009; 64: 1082–1089. 25 Wenger RH, Stiehl DP, Camenisch G. Integration of oxygen signaling at the consensus HRE. Sci STKE 2005; 2005: re12. 26 Millen J, MacLean MR, Houslay MD. Hypoxia-induced remodelling of PDE4 isoform expression and cAMP handling in human pulmonary artery smooth muscle cells. Eur J Cell Biol 2006; 85: 679–691. 27 Nunes AR, Batuce JR, Monterio EC. Functional characterization of phosphodiesterases 4 in the rat carotid body: effect of oxygen concentrations. Adv Exp Med Biol 2009; 648: 113–119. 28 Chen TC, Wadsten P, Su S, Rawlinson N, Hofman FM, Hill CK et al. The type IV phosphodiesterase inhibitor rolipram induces expression of the cell cycle inhibitors p21(Cip1) and p27(Kip1), resulting in growth inhibition, increased differentiation, and subsequent apoptosis of malignant A-172 glioma cells. Cancer Biol Ther 2002; 1: 268–276. 29 Ogawa R, Streiff MB, Bugayenko A, Kato GJ. Inhibition of PDE4 phosphodiesterase activity induces growth suppression, apoptosis, glucocorticoid sensitivity, p53, and p21(WAF1/CIP1) proteins in human acute lymphoblastic leukemia cells. Blood 2002; 99: 3390–3397. 30 Dastidar SG, Rajagopal D, Ray A. Therapeutic benefit of PDE4 inhibitors in inflammatory diseases. Curr Opin Investig Drugs 2007; 8: 364–372. 31 Lipworth BJ. Phosphodiesterase-4 inhibitors for asthma and chronic obstructive pulmonary disease. Lancet 2005; 365: 167–175. 32 Cox G, Walker RA, Andi A, Steward WP, O’Byrne KJ. Prognostic significance of platelet and microvessel counts in operable non-small cell lung cancer. Lung Cancer 2000; 29: 169–177. 33 Malec V, Gottschald OR, Li S, Rose F, Seeger W, Hanze J. HIF-1 alpha signaling is augmented during intermittent hypoxia by induction of the Nrf2 pathway in NOX1-expressing adenocarcinoma A549 cells. Free Radic Biol Med 2010; 48: 1626–1635. 34 Savai R, Schermuly RT, Voswinckel R, Renigunta A, Reichmann B, Eul B et al. HIF1alpha attenuates tumor growth in spite of augmented vascularization in an A549 adenocarcinoma mouse model. Int J Oncol 2005; 27: 393–400.

Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc)

Oncogene (2013) 1121 – 1134
