|Home | About | Journals | Submit | Contact Us | Français|
Hypoxia changes the responses of cancer cells to many chemotherapy agents, resulting in chemoresistance. The underlying molecular mechanism of hypoxia-induced drug resistance remains unclear. Pim-1 is a survival kinase which phosphorylates Bad at serine 112 to antagonize drug induced apoptosis. Here we show that hypoxia increases Pim-1 in a HIF-1α-independent manner. Inhibition of Pim-1 function by dominant negative Pim-1 dramatically restores the drug sensitivity to apoptosis induced by chemotherapy under hypoxic conditions in both in vitro and in vivo tumor models. Introduction of siRNAs for Pim-1 also resensitizes cancer cells to chemotherapy drugs under hypoxic conditions, while forced over-expression of Pim-1 endowes solid tumor cells with resistance to cisplatin, even under normoxia. Dominant negative Pim-1 prevents a decrease in mitochondrial transmembrane potential in solid tumor cells, which is normally induced by CDDP, followed by the reduced activity of Caspase-3 and -9, indicating that Pim-1 participates in hypoxia-induced drug resistance through the stabilization of mitochondrial transmembrane potential. Our results demonstrate that Pim-1 is a pivotal regulator involved in hypoxia-induced chemoresistance. Targeting Pim-1 may improve the chemotherapeutic strategy for solid tumors.
Hypoxic regions are often present in solid tumors (Koh et al., 1992; Vaupel et al., 1991). Cancer cells in hypoxic regions present a clinical challenge because they are more resistant to radiotherapy and many chemotherapeutic drugs when compared to cancer cells in normoxic regions (Gatenby et al., 1988; Teicher, 1994; Teicher et al., 1981). Irregularity in tumor blood flow causes intermittent hypoxia, which promotes solid tumor growth, angiogenesis, radioresistance and chemoresistance (Toffoli et al., 2008). The biological pathways that are regulated by hypoxia can affect apoptosis, cell cycle, angiogenesis, pH regulation, glucose uptake and glycolysis (Dewhirst et al., 2008). Hypoxia-induced chemoresistance is an important obstacle to effective cancer chemotherapy.
Hypoxia can upregulate expression of hypoxia-inducible genes by regulating the protein level of the alpha subunit of the hypoxia-inducible factor (HIF-1α), part of the heterodimeric transcription factor HIF-1. Many of these HIF-1-inducible genes directly or indirectly mediate chemoresistance, such as VEGF, Glut-1, MDR and Bcl-2 (Comerford et al., 2002; Liu et al., 2008). However, it is unlikely that the cellular response to hypoxia is mediated solely through HIF-1 (Arsham et al., 2004; Mizukami et al., 2004). PI3K/Akt signaling can induce angiogenesis and oncogenic transition both in an HIF-1-dependent and -independent manner (Arsham et al., 2004). The mechanism of hypoxia mediated chemoresistance, particularly an HIF-1α-independent manner, has not been characterized in detail. Therefore we searched for novel hypoxia-inducible genes with the use of a cDNA microarray system, comparing mRNA expression in cells cultured under hypoxia and normoxia (Niizeki et al., 2002). We have found that Pim-1 kinase expression is enhanced under hypoxia when compared with normoxia in our preliminary studies (Teh, 2004).
The oncogene, Pim-1, codes for a serine/threonine kinase (Selten et al., 1986) which is a member of the Pim kinase family that regulates apoptosis, cell cycle and gene transcription by phosphorylating target proteins such as Bad, p21/waf1, c-Myb, and c-Myc (Aho et al., 2004; Wang et al., 2002; Winn et al., 2003; Yan et al., 2003; Zippo et al., 2004; Zippo et al., 2007). Expression of Pim-1 has been shown to correlate significantly with clinical outcome in prostate cancer (Valdman et al., 2004). Pim-1 is expressed in immature hematopoietic cells, and was first identified as a common integration site in MoMuLV-induced murine T-cell lymphomas (Cuypers et al., 1984). Pim-1 enhances the survival of hematopoietic cells, in part through a Bcl-2-dependent pathway (Lilly et al., 1997; Lilly et al., 1999) and its overexpression leads to an increased incidence of hematopoietic malignancies (van Lohuizen et al., 1989). Recently it has become evident that Pim kinases may also have important roles outside the hematopoietic system (Eichmann et al., 2000). The molecular mechanism regulating Pim-1 expression is still not clear and its role of Pim-1 in solid tumor needs to be investigated in detail.
In this study, we investigate the role of Pim-1 in hypoxia-induced chemoresistance in solid malignancies. Here we show that Pim-1 is induced by hypoxia in a HIF-1α-independent manner, and plays an important role in drug resistance. Antagonizing the activity of Pim-1 using a dominant negative mutant can resensitize cancer cells to drug-induced apoptosis in cell culture and in a xenograft mouse model. Together, our studies indicate that Pim-1 is an important regulator of apoptosis which mediates chemoresistance under hypoxic conditions. Blocking the upregulation of Pim-1 by hypoxia may yield improved chemotherapies for solid malignancies.
To investigate the impact of hypoxia on the drug sensitivity of cancer cells, we cultured pancreatic cancer cell line PCI-43 under hypoxic or normoxic conditions and monitored cell viability after treatment with gemcitabine and doxorubicin by MTS assay (Figure 1a, b). PCI-43 cells were more resistant to gemcitabine (IC50: 3.24 μM compared with 0.21 μM) and doxorubicin (IC50: 1.25 μM compared with 0.29 μM) under hypoxic conditions when compared to normoxia. We also examined cell viability in PCI-43, HeLa and KMP-4 cell lines following treatment with cisplatin (CDDP) by MTS assay (Table 1). All three cell lines were more resistant to CDDP under hypoxic conditions compared to normoxic conditions. Next, we examined CDDP-induced apoptosis in a time-dependent manner in PCI-43 cells (Figure 1c and Supplemental Figure 1a). We found that hypoxia compromised the efficacy of CDDP-mediated apoptosis in these cells when compared to normoxic conditions. We also performed cell cycle analysis by PI staining and found that the percentages of sub G1 cells were decreased under hypoxia versus normoxia (Figure 1d and Supplemental Figure 1b). Moreover, we examined CDDP-induced apoptosis by Annexin V staining in three more cancer cell lines (KMP-4, HeLa and PCI-35) under hypoxic or normoxic conditions, and found that CDDP-induced apoptosis was reduced by approximately 50% under hypoxia compared to normoxia (Figure 1e and Supplemental Figure 1c). Interestingly, hypoxia did not have an impact on anti-Fas-mediated apoptosis in PCI-43 cells as well as in HeLa cells (Supplemental Figure 2). We also observed that hypoxia conferred resistance to fluorouracil and A23187, which are not substrates of the multiple drug resistance protein (MDR) in PCI-43 cells (Figure 1f and Supplemental Figure 1d), as well as in HeLa cells (Figure 1g and Supplemental Figure 1e). Thus, we conclude that hypoxic conditions lead to chemotherapy resistance.
Previously, with the use of a cDNA micro-array system, it was found that Pim-1 was expressed at a higher level under hypoxia than under normoxia in pancreatic cancer cell line PCI-10 (Niizeki et al., 2002). Since Pim-1 is an anti-apoptotic factor in hematopoietic cells (Lilly et al., 1999; van Lohuizen et al., 1989), we hypothesized that Pim-1 might also be an anti-apoptotic factor induced by hypoxia in solid tumor cells. Indeed, a panel of human cancer cell lines expressed higher levels of Pim-1 protein, as well as mRNA, under hypoxia than under normoxia (Figure 2a).
We then examined the expression of Bcl-2 family proteins as well as another Pim family protein, Pim-2, under hypoxic and normoxic conditions. Hypoxic conditions did not change the expression level of Bcl-2, Bad, Bid, Bcl-xL, Bax or Pim-2 in PCI-43 cells (Fig. 2b); however, Bad protein was phosphorylated after a 1-hour exposure to hypoxia. Since Bad is a legitimate substrate of Pim-1 and Pim-1 enhances Bcl-2 activity by phosphorylating Bad (Aho et al., 2004), we hypothesized that Pim-1 might be an earlier hypoxia-inducible target, paralleling Bad phosphorylation under hypoxia. We further characterized the kinetics of induction of Pim-1 by hypoxia and found that the time to maximum induction of Pim-1 (about 1 hour) preceded the time to maximum induction of HIF-1α (about 6 hours), which is known to upregulate the expression of VEGF and Glut-1 (Figure 2c). Therefore, it was very unlikely that hypoxia increased Pim-1 via HIF-1α in the same mechanism as typical targets of HIF-1α, such as VEGF and Glut-1. Furthermore, we checked the expression of these proteins in a dominant negative HIF-1α stable transfectant, dnHIF-1α, which lacks the DNA-binding, transactivation and oxygen dependent degradation domains. dnHIF-1α is reported to act in a dominant negative manner through the inhibition of functional HIF-1 heterodimer formation without HIF-1α protein level changes (Chen et al., 2003). Figure 2c also shows that the expression of Pim-1 was still induced by hypoxia in cells transfected with dnHIF-1α while the expression of VEGF and Glut-1 were significantly suppressed by dnHIF-1α. Also, we found that Pim-1 is still induced under hypoxic condition even in HIF-1α knock-down HCT116 and HeLa cells (Figure 2d). Taken together, these results suggest that Pim-1 is regulated by hypoxia in a HIF-1-independent manner.
To further confirm whether hypoxia activated Pim-1 transcription through interaction of HIF-1 with the Pim-1 promoter, we employed the electrophoretic mobility shift assay (EMSA). The Pim-1 promoter contains 4 putative hypoxia response elements (HRE regions [DNA consensus binding motif: 5′-RCGTG-3′] ) (Comerford et al., 2002; Semenza, 1999) located at positions: −315~−306 (HRE1), −768~−759 (HRE2), −874~−865 (HRE3) and −965~−956 (HRE4) (Supplemental Figure 3). While HIF-1 did not bind to any of the four Pim-1 promoter probes (Figure 2e), more HIF-1 bound to the positive control EPOwt33 probe under hypoxia than under normoxia, and this increase in binding was reversible by competition with unlabeled EPOwt33 probe. These results suggest that the putative HIF-1 binding sites in the Pim-1 promoter region are not being utilized by HIF-1.
Next, we cloned the Pim-1 promoter (−1715 to +38bp) into a luciferase reporter vector, pGL3-basic plasmid. As shown in Figure 2f, no increase in luciferase activity was found in 293 cells co-transfected with pGL3-proPim-1 and HIF-1α plasmids under normoxia, when compared to those transfected with pGL3-proPim-1 alone. However, in the positive control, luciferase activity was increased in cells co-transfected pGL3-5×HRE with HIF-1α plasmids under normoxia. Interestingly, luciferase activity was increased in the cells transfected with pGL3-proPim-1 under hypoxia, suggesting that hypoxia can enhance Pim-1 expression in a HIF-1-independent manner, probably through other hypoxic pathways.
To define the role of Pim-1 in chemoresistance under hypoxia, we antagonized the function of Pim-1 by transfecting PCI-43 cells with dominant-negative Pim-1 (dnPim-1). dnPim-1, which inhibits the function of wild type Pim-1 (Lilly et al., 1999), lacks a kinase activation domain (1–81aa) (Figure 3a). In Figure 3b, we evaluated the expression of dnPim-1 in three stable dnPim-1-transfectants. Since Pim-1 is a survival kinase that is known to phosphorylate Bad at ser112, which is an important inactivation site, thus promoting cell survival, we confirmed the antagonistic effect of dnPim-1 on Pim-1 mediated phosphorylation of Bad (Figure 3c). The phosphorylation of Bad at ser112 was also increased under hypoxia in vector control but not in dnPim-1 transfectants. There were no changes in the total protein levels of Bad (Figure 3c). These results indicate that hypoxia increases Pim-1 which phosphorylates and inactivates Bad to inhibit apoptosis. Next, we examined the sensitivity of dnPim-1 transfectants to apoptosis. As expected, dnPim-1 transfectants and vector control cells had small percentages of apoptotic cells in the absence of drug treatment (Figure 3d and Supplemental Figure 4). Hypoxia inhibited CDDP-induced apoptosis in vector-transfected control cells (V3), but this hypoxia-induced chemoresistance was no longer present in dnPim-1 transfectants (dnP3 and dnP4) (Figure 3e and Supplemental Figure 4). Therefore, dnPim-1 expression resensitized cells to CDDP under hypoxia. In contrast, hypoxia did not antagonize anti-Fas Ab-induced apoptosis (Figure 3f and Supplemental Figure 2) and transfection of dnPim-1 did not influence this effect (Figure 3f and Supplemental Figure 4). These results indicate that dnPim-1 effectively suppresses Pim-1 function, and resensitizes cells to CDDP under hypoxic conditions.
To further confirm the role of Pim-1 in hypoxia-induced chemoresistance, we also employed molecular methods to knockdown Pim-1 levels. Under hypoxia, siRNAs for Pim-1 effectively suppressed Pim-1 protein expression in PCI-43 cells (Figure 4a), and siRNA suppression of Pim-1 restored sensitivity to CDDP (Figure 4b and 4c). Conversely, we established Pim-1-overexpressing stable transfectant PCI-43 cells (Figure 4d) and found that over-expression of Pim-1 rendered the cells resistant to CDDP under normoxia (Figure. 4e and 4f), underscoring Pim-1’s role as a mediator of chemoresistance.
To examine the sensitivity of dnPim-1 transfectants to CDDP in vivo, we used Tet-On dominant negative transfectants of HeLa cells, which expressed dnPim-1 protein in the presence of doxycycline (Figure 5a). Administration of doxycycline together with CDDP completely suppressed the growth of the Tet-On dnPim-1-transfectants in a SCID mouse xenograft model. In contrast, CDDP could only slightly suppress the growth of the transfectants when dnPim-1 was not induced (Figure 5b). Since most solid tumors contain hypoxic regions, these results suggest that Pim-1 induced by hypoxia may be responsible for the intrinsic resistance to anti-cancer drugs in solid tumor cells and that the suppression of Pim-1 function could be helpful for the chemotherapy of solid tumors.
Since Pim-1 had no effect on apoptosis induced by anti-Fas antibody (Figure 3f), we hypothesized that Pim-1 might induce resistance to anti-cancer drugs by inhibiting the intrinsic mitochondrial apoptosis pathway through stabilization of MTP. We used TMRE (an indicator of mitochondrial transmembrane potential) to measure MTP by FACS (Figure 6a and 6b). We found that the MTP in vector control cells after exposure to CDDP for 48 hours was higher under hypoxia than under normoxia (Figure 6a, peak fluorescence indicated by arrows). By contrast, there is no difference in MTP in dnPim-1 transfectants between hypoxia and normoxia at the same time point. Consistent with the above findings, CDDP-induced Caspase-9 and Caspase-3 activities under normoxia were compromised by hypoxia in control cells (Figure 6c and 6d). As for the dnPim-1 transfectant, Caspase-9 and Caspase-3 activities were increased by CDDP regardless of hypoxia or normoxia (Figure 6c and 6d) suggesting again that inhibiting Pim-1 activity resensitized cells to chemotherapy. Note that the anti-Fas antibody had no impact on Caspase-9 and Caspase-3 activities in either control cells or dnPim-1 transfectants (Figure 6c and 6d). The specificities of the Caspase-9 and Caspase-3 assays were confirmed using Z-LEHD-FMK (Caspase-9 inhibitor) and Ac-DEVD-CHO (Caspase-3 inhibitor), respectively. CDDP did not increase the activity of Caspase-8 while anti-Fas Ab increased Caspase-8 activity regardless of normoxia or hypoxia (Figure 6e). Ac-IETD-CHO (Caspase-8 inhibitor) was used to confirm the specificity of the Caspase-8 assay. Together, these data indicate that Pim-1 mediated hypoxia-induced chemoresistance by affecting MTP and regulating the activities of Caspase-9 and Caspase-3.
In this study we show that hypoxia causes resistance to chemotherapy in a variety of cancer cells. The underlying molecular mechanism of hypoxia-induced chemoresistance has not been previously characterized in detail. Here, we show that Pim-1 can be induced by hypoxia and plays an important role in drug resistance to CDDP, gemcitabine, and fluorouracil. The phosphorylation of Bad at ser112 by Pim-1, connects hypoxia to the inhibition of apoptosis, offering a novel mechanism as to how hypoxia contributes to compromised drug therapy response.
HIF-1 plays an important role in hypoxia-induced gene expression. Scanning the 5′-promoter region of Pim-1 for the consensus sequence of the HIF-1 response element (HRE) has revealed several potential binding sites for HIF-1. We have tested the possibility that HIF-1 can regulate the expression of Pim-1 by binding to these putative HREs and found this is not the case. Three lines of evidence demonstrate that the activity of HIF-1 is not a critical regulator of Pim-1. First, HIF-1 does not bind to the putative HRE sites in the Pim-1 promoter region as assayed by EMSA (Figure 2e). Second, HIF-1α cannot enhance luciferase activity in pGL3-proPim-1 reporter, which contains the Pim-1 promoter sequence (−1715 to +38bp) (Figure 2f). Last, in dominant negative HIF-1α as well as in HIF-1α knockdown cells, inhibition of HIF-1α expression and activity cannot block the increase of Pim-1 protein level under hypoxia (Figure 2c and 2d). These results strongly suggest that the mechanism of up-regulation of Pim-1 transcription is independent of HIF-1α. Since some previous studies demonstrated that AP-1 and NF-κB were involved in the up-regulation of various hypoxia-inducible genes (Bandyopadhyay et al., 1995; Koong et al., 1994), it is possible that these transcription factors may be involved in the regulation of Pim-1 under hypoxia. Moreover, rapid and prominent elevation of the Pim-1 protein, earlier than HIF-1 appearance, after exposure to hypoxia (Figure 2b), favors the notion that the expression of Pim-1 protein may be regulated post-transcriptionally, in a HIF-1-independent manner (Qian et al., 2005).
Pim-1 protein is induced earlier than other hypoxia-induced factors, suggesting that Pim-1 displays its role in survival of solid tumor cells following exposure to acute, as well as, chronic hypoxia. Most other hypoxia-inducible factors, such as, VEGF, glucose transporters and multidrug resistance 1 (MDR1), are induced in an HIF-1-dependent manner when exposed to chronic hypoxia. It is described that tumor hypoxia develops as a result of two independent phenomena: chronic hypoxia caused by limitation of oxygen diffusion, and transient hypoxia caused by instability of microvessel flow (Cardenas-Navia et al., 2004). In our in vivo result, a remarkable regression of tumors (Figure 5) obtained by the combination of CDDP and dnPim-1, can be explained by our recent observation that dominant negative Pim-1 also has an impact on tumor angiogenesis by significantly decreaing CD31 expression in tumor tissue (Chen et al., 2009). Few studies have reported on the role of Pim-1 in vasculogenesis or angiogenesis. These results, in combination with a recent report that demonstrated Pim-1 is required for VEGF-A-dependent proliferation and migration of endothelial cells (Zippo et al., 2004), implies that Pim-1 may modulate VEGF-induced angiogenesis.
Pim-1 is a survival kinase and phosphorylates many targets that regulate apoptosis, cell cycle and gene transcription. How Pim-1 exerts its influence on cells in preventing them from undergoing apoptosis remains unclear. Apoptosis is accompanied by a variety of mitochondrial dysfunctions. For instance, loss of MTP and production of reactive oxygen species (ROS) are early events in cells destined to undergo apoptosis (Zamzami et al., 1995). Pim-1 is considered to inhibit the production of ROS indirectly, during anti-cancer drug-induced apoptosis (Lilly et al., 1999), resulting in the restoration of MTP. Bad is a legitimate substrate of Pim-1, and Pim-1 can enhance Bcl-2 activity by phosphorylating Bad. In the present study, Bad was phosphorylated under hypoxia, while no phosphorylation of Bad was observed in dominant negative Pim-1 transfectants under normoxia or hypoxia. These findings suggest that the phosphorylation of Bad may be responsible for the drug sensitivity decrease induced by overexpression of Pim-1 under hypoxic conditions. To determine whether the hypoxia-Pim-1 axis is also regulating the death receptor pathway to affect hypoxia-mediated drug resistance, we checked the expression of Bid, which is a Bcl-2 family protein, interposed between the death receptor pathway (Fas and TNF) and the mitochondrial pathway in the vast majority of Fas-sensitive cells. Although Bid is known to translocate itself to the mitochondria, causing pro-apoptotic changes (Luo et al., 1998). We found that Bid was not involved under hypoxia (Figure 2b). Also, our observations showed that anti-Fas antibody induction of Caspase-8 and the resulting apoptosis were not affected by hypoxia (Figure 6d). Taken together, these results suggest that hypoxia-induced Pim-1 signaling is related to mitochondrial dysfunction (loss of MTP) but independent of the Fas-Caspase-8-Bid pathway.
Given that hypoxia is associated with chemoresistance and that hypoxia can enhance Pim-1 expression, it is conceivable that Pim-1 would be an ideal target for rational cancer therapy. An important new concept advanced here is that antagonizing Pim-1 activity can sensitize cancer cells to chemotherapy drug and justifies Pim-1 as an important molecular target for developing specific small molecule inhibitors for combination chemotherapy, and such drug development effort has begun (Pierce et al., 2008).
Pancreatic ductal adenocarcinoma cell lines (PCI-10, PCI-35, PCI-43 cells and KMP-4) were kindly supplied by Dr. Hiroshi Ishikura (The First Department of Pathology, Hokkaido University School of Medicine) and Dr. Yutaka Shimada (Department of Surgery & Surgical Basic Science Graduate School of Medicine, Kyoto University). Dominant negative HIF-1α transfectants were established by Dr. Jian Chen et al. (Chen et al., 2003). Annexin-V-FLUOS staining kit was purchased from Japan Roche Diagnostic Co. Ltd. (Tokyo, Japan). Small interfering RNAs (siRNAs) for Pim-1 were synthesized with the Silencer™ siRNA Construction kit (Ambion, Cat. 1620). Sequences for Pim-1 siRNA are shown in Supplemental Table 1. siRNAs for HIF-1α (ON-TARGETplus SMARTpool siRNA) and siRNA control (ON-TARGETplus Non-targeting Pool), were purchased from Thermo Scientific Dharmacon (Chicago, IL). Sequences for HIF-1α siRNA are shown in Supplemental Table 1. Caspase activity assay kits for Caspases-3,-8, and-9 were purchased from BD Biosciences (San Jose, CA). Inhibitors of Caspase-3,-8, and-9 (Ac-DEVD-CHO, Ac-IETD-CHO, Z-LEHD-FMK) were purchased from BD Biosciences (San Jose, CA). Carbonylcyanide m-chlorophenylhydrazone (CCCP) was purchased from Sigma-Aldrich Japan (Tokyo, Japan).
Sensitivity of tumor cell lines to anti-cancer drugs was determined by a colorimetric MTS assay according to the manufacturer’s instructions (Cell Titer 96 Aqueous Non-radioactive Cell Proliferation Assay; Promega; Madison, WI). 2 ×103 cells were incubated with different doses of anti-cancer drugs for 48 h under normoxia (20 % O2/5 % CO2) or hypoxia (1% O2/5 % CO2). Incubation under hypoxia was done in a hypoxic chamber gassed with 95 % N2 and 5 % CO2 (Wakenyaku Co. Ltd., Tokyo). After incubation, the cells were incubated for 3 h with MTS solution. Subsequently, the absorbance was measured at 490 nm on an ELISA reader (BioRad Model 550).
Sensitivity of tumor cell lines to apoptosis was determined by two-color analysis using propidium iodide (PI) and FITC-conjugated anti-annexin V according to the manufacturer’s instructions. After incubation with or without CDDP, at 25 μM for 48 h, under normoxia or hypoxia, the cells were stained with PI and FITC-conjugated anti-annexin V and analyzed with a FACScalibur flow cytometer (Becton Dickinson, Mountain View, CA).
Cell cycle analysis was performed with propidium iodide (PI) staining. After incubation with or without CDDP, at 50 μM for 72 hours, under normoxia or hypoxia, the cells were washed twice with PBS, harvested by trypsinization, fixed with ice-cold 70% ethanol, treated with RNaseA (0.25mg/ml), and stained with PI (0.02 mg/ml). Analyses were performed with a FACScalibur flow cytometer (Becton Dickinson, Mountain View, CA).
Total proteins were separated on 4–12 % gradient polyacrylamide gels and electro-transferred to PVDF membranes. Membranes were incubated with anti-Pim-1 antibody (Transduction Laboratories, Inc., Lexington, KY, USA), anti-Pim-2 antibody (Santa Cruz Biotechnology, Inc., Santa Cruz, CA ), anti-Bcl-2 antibody (Pharmingen, Inc., San Diego, CA), anti-Bax antibody (Pharmingen), anti-Bcl-xL antibody (Cell Signaling Technology, Inc.), anti-Bid antibody (Santa Cruz Biotechnology), anti-Bad antibody (Cell Signaling Technology), anti-Phospho-Bad (Ser112) antibody (Cell Signaling Technology), anti-HIF-1α antibody (Santa Cruz Biotechnology), anti-VEGF antibody (Up-state Biotechnology, Inc., Lake Placid, NY), anti-Glut-1 antibody (Immuno-Biological Laboratories Inc., Gunma, Japan), anti-HIF-1α antibody (Novus Biologicals Inc., Littleton, CO), anti-HA antibody (Santa Cruz Biotechnology), anti-FLAG M2 monoclonal antibody (Sigma, Saint Louis, USA) or anti-actin antibody (Santa Cruz Biotechnology).
Northern blot analysis was performed by the method described previously (Chen et al., 2003). Total RNA (25 mg) was separated by electrophoresis in 1% denaturing formaldehydeagarose gels. The RNA was transferred to a nylon membrane (Hybond, Amersham) by capillary elution overnight and UV cross-linked. The membrane was hybridized overnight at 42°C with the cDNA probe labeled with 32P by the use of a random primer DNA labeling kit (Takara Biomedicals, Tokyo, Japan) for Pim-1. Probes for Pim-1 were obtained by PCR amplification with the use of primers as follows: Pim-1 forward, 5′-ggttggatgctcttgtccaa-3′; reverse, 5′-ccttccagaagtcttctat-3′.
Total RNA was extracted with the use of TRIZOL reagent according to the manufacturer’s instructions. Reverse Transcription were performed by using SuperScriptn III First-Strand kit (Invitrogen). Each cDNA (10ng) was amplified in triplicate with iQ SYBR Green Supermix PCR kit (BIO-RAD) for 40 cycles on a Bio-Rad system (iCycler Thermal Cycler). Sequences for Pim-1 primer and β-actin primers are shown in Supplemental Table 1.
EMSA was performed according to the previously described method (Chen et al., 2003). After 16 h incubation of PCI-43 cells in hypoxia and normoxia, nuclear extracts were obtained. Double-stranded HIF-1-specific EPO wt33 oligonucleotide probe containing two HIF-1-binding sites in tandem (5′-gccctacgtgctgtctcacacagcctgtctga-3′ and 5′-gtcagacaggctgtggagacagcacgtaggg-3′) were end-labeled with [32P] dCTP by Klenow fragment. Nuclear extract (3 μg) was incubated with a 300 fmol probe in 30 ml binding buffer (10 mM HEPES, pH 7.8, 50 mMKCl, 1 mM ethylenediaminetetraacetic acid, 5 mM MgCl2,10% v/v glycerol, and 2 μg of poly-dI-dC) for 20 minutes at room temperature. For the competition assay, a 50-fold molar excess of unlabeled oligonucleotide probe was added to nuclear extracts for 15 minutes before the addition of labeled probe. Sequences for synthetic oligonucleotide probes are shown in Supplemental Table 2.
The Pim-1 promoter region was amplified by PCR with primers: 5′-GCTGTATGACCCTGTGCCAT-3′ sense, 5′-CTGCAGCAGACGCCCGGCTC -3′ antisense, and ligated into pGL3-basic vector. 293 cells were seeded in 24-well plates at 5 × 104 cells per well overnight, and then were co-transfected with plasmids pGL3-proPim-1 (0.5 μg) expressing the sequence corresponding to proPim-1 (−1715 to +38bp) of the Pim-1 promoter, or pGL3-5×HRE (0.5 μg) which included five tandem repeats of the erythropoietin gene HRE sequence (5′-CGCCCTACGTGCTGTCTCACACAGCCTGTCTGA-3′) and pcDNA3.1-HIF-1α (0.5μg) expression vector with the Lipofectamine™ 2000 transfection kit. The Rapid Response™ phRL-CMV Vector (Catalog No. E6271, Promega) (50 ng) was transfected simultaneously and used as an internal transfection control. In subsets of experiments, cells were transfected with empty pGL3-basic vector (Promega) to control background luciferase activity. After transfection, cells were subjected to hypoxia or normoxia for 16 h. Luciferase activity was assessed using a Dual-Luciferase® Reporter assay system (Promega) and Lumat LB 9506 luminometer (Berthold) according to the manufacturer’s instructions.
cDNA for dominant negative Pim-1, which lacks the kinase activation domain (Lilly et al., 1999), was amplified from reverse transcription products of mRNAs purified from PCI-10 cells and cloned into PCR4-TOPO. Plasmids were sequenced with a DyeDeoxy Terminator kit (Perkin-Elmer, Urayasu, Japan) on an ABI 377 automated sequencer (Applied Biosystems, Urayasu, Japan) according to the manufacturer’s protocol. Cloned fragments were ligated into PcDNA3.1+ (Invitrogen, Carlsbad, CA). PCI-43 cells were transfected with the expression vector with the use of Lipofectamine (Life Technologies, Tokyo, Japan). Transfectants were cloned by a limiting dilution method following selection with G-418 (1,200 mg/ml). Transfectants were maintained in the presence of 600 mg/ml of G-418. Sequences of PCR primers are shown in Supplemental Table 1.
The mitochondria-specific dye, Tetramethylrhodamine ethylester perchlorate (TMRE) (200 nM) (Molecular Probes) was added during the last 30 minutes of treatment with hypoxia or normoxia in PCI-43 cells. The medium was transferred into a 75-mm Falcon polystyrene tube and the adherent cells were trypsinized and collected into the same tube. After washing with PBS, the cells were analyzed by the use of a FACScan flow cytometer(Becton Dickinson, Sunnyvale, CA) for mitochondrial uptake. Untreated cells and cells treated with 50 μM CCCP were used as negative controls and positive controls, respectively. Caspase activities were determined according to the manufacturer’s instructions. dnPim-1 transfectants and vector controls were treated with CDDP under normoxia or hypoxia for 48 hours, and cells lysed. 50 μl of cell lysate was incubated with 15 μl of each Caspase fluorogenic substrate (DEVD-AFC for Caspase-3; IETD-AFC for Caspase-8; LEHD-AMC for Caspase-9) for 1 hour at 37°C. Subsequently, the amount of AFC or AMC liberated from substrates was measured by a spectrofluorometer. The inhibitors of Caspase-3, -8, and -9, (Ac-DEVD-CHO; Ac-IETD-CHO; Z-LEHD-FMK) were added in each mixture as controls.
A tetracycline-inducible expression system and the HeLa/Tet-On cell line, a HeLa clone expressing a reverse tetracycline-controlled transactivator, were purchased from BD Biosciences Clontech (Palo Alto, CA, USA). HeLa Tet-On cells were transfected with pTRE/dominant negative Pim-1 and PTK-Hyg, a plasmid expressing the hygromycin resistance gene (BD Biosciences Clontech). The cells were cultured for more than 4 weeks in medium containing hygromycin (200 mg/ml) to obtain stable transfectants.
SCID mice were housed in AAALAC approved barrier facilities on a 12-hour light/dark cycle, with food and water ad libitum. The mice were treated under approved protocols in compliance with the animal care and use guidelines in accordance with international standards including the NIH of USA. Five million cells of dnPim-1 transfectants were injected subcutaneously into the right flank of each mouse (n=5, in each group). Tumor formation was observed every three days for 3 weeks after inoculation. Tumor volumes were measured with this formula: tumor volume = 0.5 × ab2 (a, major axis; b, minor axis) (Yoshida et al., 1989). Body weight, feeding behavior and motor activity were monitored thrice weekly as indicators of general health. Animals with the following conditions were euthanized: >10% weight loss, motor retardation, inability to obtain food or water, ruffled hair, or largest diameter of the tumor >15 mm.
All statistical analyses except the analysis of tumor sizes in the xenograft study were performed using an unpaired Student’s t test for two groups and one-way analysis of variance with post-hoc intergroup comparisons using Tukey test for multiple groups. The xenograft data were analyzed by linear mixed models (SPSS for Windows, version 12.0, SPSS, Inc. Chicago, IL) with comparisons of fixed effects using the restricted maximum likelihood method.
Supplemental Figure 1. Quantification of FACS analysis for Figure 1. All statistical analyses were performed using an unpaired Student’s t test. (a) Mean ± SD apoptotic cell percentage in three independent experiments of Figure 1c. (b) Mean ± SD apoptotic cell percentage in three independent experiments of Figure 1d. (c) Mean ± SD apoptotic cell percentage in three independent experiments of Figure 1e. (d) Mean ± SD apoptotic cell percentage in three independent experiments of Figure 1f. (e) Mean ± SD apoptotic cell percentage in three independent experiments of Figure 1g. * indicates significance (P<0.01).
Supplemental Figure 2. Hypoxia does not antagonize anti-Fas antibody-induced apoptosis in tumor cells. (a) PCI-43 cells treated with anti-Fas antibody were cultured under hypoxic conditions for the indicated times and the percentage of apoptotic cells determined by PI and annexin V staining using FACS analysis. (b) Mean ± SD apoptotic cell percentage in three independent experiments of a. (c) HeLa cells treated with anti-Fas antibody were cultured under hypoxic conditions and the percentage of apoptotic cells determined as in (a). (d) Mean ± SD apoptotic cell percentage in three independent experiments of c.
Supplemental Figure 3. Sequences of Pim-1 promoter putative HREs. 4 putative HREs in Pim-1 promoter region. The HRE site in Glut-1 gene promoter is shown for comparison.
Supplemental Figure 4. Quantification of FACS analysis for Figure 3. All statistical analyses were performed using an unpaired Student’s t test. Mean ± SD apoptotic cell percentage in three independent experiments of Figure 3d, 3e and 3f. * indicates significance (P<0.01).
This work was supported in part by the NIHRO1CA (089266 Lee M H), Grants-in-Aid for Scientific Research from the Japan Society for the Promotion of Science (JSPS) (Chen J), Grants-in-Aid for Cancer Research from the Japanese Ministry of Education, Culture, Sports, Science and Technology (Kobayashi M), U. S. Department of Defense Breast Cancer Research Program of the Office of the Congressionally Directed Medical Research Programs (DOD SIDA BC062166 Yeung S J & Lee M H) and Cancer Center Core Grant (CA16672). We thank Dr. Vuong BQ for providing the pCDNA3.1-His-SOCS1 plasmid. We thank Mrs. Shirley Ware-Gully for her assistance in preparing the manuscript.
Conflict of interest
The authors declare no conflict of interest.
Supplementary information is available at Oncogene’s website