Oncogene (2001) 20, 2937 ± 2945 ã 2001 Nature Publishing Group All rights reserved 0950 ± 9232/01 $15.00 www.nature.com/onc

Adrenomedullin inhibits hypoxic cell death by upregulation of Bcl-2 in endometrial cells: a possible promotion mechanism for tumour growth

MK Oehler1,3, C Norbury2, S Hague1,3, MCP Rees3 and R Bicknell*,1

1Molecular Laboratory, Imperial Cancer Research Fund, Institute of Molecular Medicine, John Radcli€e Hospital, University of Oxford, Oxford, OX3 9DS, UK; 2Cell Cycle Group, University of Oxford, Oxford, OX3 9DS, UK; 3Nueld Department of Obstetrics and Gynaecology, University of Oxford, Oxford, OX3 9DU, UK

Regions of are a common feature of solid tumours. of benign and malignant tissues and human cancer cell When tumour cells are exposed to hypoxic stress, lines (Hinson et al., 2000). transcription of a battery of is initiated. The ADM is a pluripotent with properties angiogenic factor (ADM) is a hypoxia ranging from inducing vasorelaxation to acting as a regulated . ADM is thought to act through the G for various benign and malignant cell -coupled receptor receptor-like receptor types (Hinson et al., 2000). We were the ®rst to show (CRLR), with speci®city being conferred by the receptor that ADM is an angiogenic factor (Zhao et al., 1998). associated modifying protein 2 (RAMP2). Here we report Furthermore we demonstrated that ADM expression in for the ®rst time that ADM treated or stably transfected uterine leiomyomas (benign smooth muscle tumours) Ishikawa cells overexpressing ADM show increased correlated with vascular density (Hague et al., 2000). In resistance to hypoxia induced apoptosis. These cells also addition, we reported the presence of ADM in human show an upregulation of the oncoprotein Bcl-2, which is endometrium where it was upregulated by the non- protective against hypoxic cell death when transiently steroidal antiestrogen tamoxifen (TAM) (Zhao et al., transfected into Ishikawa cells. Since Ishikawa cells 1998). TAM, the long term endocrine treatment of express the putative ADM-receptor CRLR ± RAMP2 the choice for selected patients with breast cancer, is known production and secretion of ADM with the consecutive to induce proliferative changes of the endometrium, upregulation of Bcl-2 could establish an autocrine/ increasing the risk of developing endometrial cancer paracrine mechanism rescuing malignant cells from (Bergman et al., 2000; Carthew et al., 2000). Recent hypoxic cell death. These results, taken together with our evidence of a role for ADM in vascular development previous ®ndings that ADM is an angiogenic factor which has come from the ADM knockout mouse that shows is upregulated by the nonsteroidal antiestrogen tamoxifen extreme hydrops fetalis and cardiovascular abnormal- (TAM) in endometrial cells, implicate this peptide as a ities (Caron and Smithies, 2001). promoter of tumour growth and a possible target for Recently, ADM has been reported to be upregulated anticancer strategies. Oncogene (2001) 20, 2937 ± 2945. by hypoxia (Cormier-Regard et al., 1998; Nakayama et al., 1998). Expression was shown to be under hypoxia- Keywords: adrenomedullin; Bcl-2; hypoxia; apoptosis; inducible transcription factor-1 (HIF-1) regulation cell death; survival similar to that of classic oxygen-regulated genes such as (Epo) or vascular endothelial growth factor (VEGF) (Garayoa et al., 2000; Nguyen and Introduction Claycomb, 1999). However, the function of ADM in hypoxic cells is currently unknown. Adrenomedullin (ADM) is a 52-amino-acid peptide Regions of hypoxia are a common feature of solid which belongs to the calcitonin gene-related peptide tumours. Uncontrolled malignant growth leads to family (Wimalawansa, 1997). ADM has been shown to limited vascular supply and consequently to hypoxia act through the G protein-coupled receptor calcitonin at distances beyond the di€usion capacity of oxygen receptor-like receptor (CRLR), with a speci®city for (100 ± 150 mm) (Thomlinson and Gray, 1955). It is ADM being conferred by the receptor associated thought that hypoxia functions as a selective pressure modifying protein 2 (RAMP2) (Kamitani et al., 1999; for solid tumours leading to progression to a more McLatchie et al., 1998). ADM, initially isolated from a malignant phenotype by selection for an apoptosis- human phaeochromocytoma, is expressed in a variety resistant cell population (Graeber et al., 1996). Apoptosis (also called programmed cell death) is a genetically controlled, morphologically unique process, which plays a central role in regulating tissue home- *Correspondence: R Bicknell Received 1 November 2000; revised 26 February 2001; accepted 26 ostasis by eliminating cells that are harmful or no February 2001 longer required. Dysregulation of apoptosis results in Adrenomedullin in tumourogenesis MK Oehler et al. 2938 abnormal cell growth and death, and can promote tumour expansion (Naik et al., 1996). Bcl-2 is the prototype of a class of oncogenes which is involved in the regulation of apoptosis (Adams and Cory, 1998; Pellegrini and Strasser, 1999). Bcl-2 acts as an anti- apoptotic factor which suppresses programmed cell death triggered by stimuli such as hypoxia (Shimizu et al., 1995). We have studied the role of ADM in endometrial carcinoma cells under hypoxia. Our data indicate for the ®rst time that ADM confers resistance to hypoxic cell death in an autocrine/paracrine manner. This anti- apoptotic action appears to be mediated by upregula- tion of the Bcl-2 oncoprotein. These results, taken together with our previous ®ndings that ADM is an angiogenic factor which is upregulated by tamoxifen, implicates ADM as a promoter of tumour growth in human endometrium.

Results

ADM mRNA and peptide expression is increased under hypoxia RNAse protection assay showed a time dependent upregulation of ADM mRNA expression under hypoxia. ADM mRNA expression levels after 18 h of hypoxia were about 15-fold higher compared with the normoxic controls (Figure 1a). This result could also be mimicked with cobalt chloride (500 mM), which induced a 10-fold increased expression of ADM mRNA after an incubation period of 8 h (data not shown). We conducted further experiments to address the question of whether the induction of ADM mRNA under hypoxic conditions is accompanied by an increase of peptide expression. Since ADM is known to be a peptide which is rapidly secreted from malignant cells, concentrations in the conditioned media of Ishikawa cells under normoxia and hypoxia Figure 1 E€ect of hypoxia on ADM mRNA and peptide were evaluated (Miller et al., 1996). Values of expression. (a) Ishikawa cells were exposed to 0.1% O2 for up accumulated IR-ADM in cultured media of cells under to 18 h and total RNA isolated at di€erent time points. ADM 0.1% O were signi®cantly higher than those from mRNA expression was analysed by RNAse protection assay. (b) 2 Supernatant from Ishikawa cells after 24 h exposure to hypoxia normoxic cells after 24 h (P50.05) (Figure 1b). or normoxia were analysed for accumulated ADM-peptide by RIA. Data represent the means of three replicates, with bars indicating s.e.m. Ishikawa-cells express the CRLR and RAMP2 We asked whether Ishikawa cells could be capable of responding to the secreted ADM under hypoxia. RT ± PCR was used to evaluate the expression of the cell death (sub-G1 peak) was observed after 36 h of putative ADM-receptor CRLR ± RAMP2. Ishikawa hypoxia at a concentration of 1076 M ADM (P50.05) cells both expressed CRLR and RAMP2 (Figure 2). (Figure 3). This result points to a possible autocrine and/or To distinguish apoptotic from necrotic cells, double paracrine mechanism of action of ADM. staining for exposed phosphatidylserine (Annexin-V) and propidium iodide (PI) exclusion was performed using un®xed cells. At a concentration of 1076 M ADM Addition of ADM increases hypoxic cell survival a signi®cant reduction of early apoptotic cells (Annexin In the hypoxia experiments addition of ADM dose V-positive cells) and late apoptotic/necrotic cells dependently (10712 ±1076 M) inhibited hypoxia induced (Annexin-V-positive and PI-positive cells) was observed cell death in Ishikawa cells. About 40% less hypoxic after 36 h at 0.1% O2 (P50.05) (Figure 4).

Oncogene Adrenomedullin in tumourogenesis MK Oehler et al. 2939

Figure 2 Expression of CRLR and RAMP2 by Ishikawa cells detected by RT ± PCR analysis. The presence of the 222/226 bp band indicates expression of CRLR and RAMP2 respectively. The left hand lane is molecular weight markers

Stably transfected ADM-overexpressing Ishikawa cells are more resistant to hypoxic cell death To further evaluate the ability of ADM to protect Ishikawa cells from hypoxic cell death, stable transfec- tants were created. Two Ishikawa cell clones (Ishi- ADM1 and Ishi-ADM2) highly expressing ADM mRNA and peptide were identi®ed by RT ± PCR, immunohistochemistry and FACS analysis (Figure 5). Those clones were examined for thir resistance to hypoxic cell death. After exposure to hypoxia the sub- G1 populations of both ADM high-expressing cell clones were up to 30% lower than those of control clones (P50.05) (data not shown). Staining for Annexin-V also revealed a signi®cant increased survival of transfected cells vs controls (P50.05) (Figure 6). Figure 3 ADM enhances tumour cell survival under hypoxia. (a) Ishikawa cells were exposed to 36 h 0.1% O2 with di€erent concentrations of ADM. The proportion of cells undergoing cell ADM increases Bcl-2 expression death in each population was estimated by detergent extraction of In order to elucidate the mechanism underlying the the cells followed by PI staining and quanti®cation of the sub- diploid cells (sub-G1 peak) using a FACScan. (b) Data represent suppressive e€ect of ADM on hypoxic cell death we the means of ®ve independent experiments, with bars showing examined whether ADM might alter the expression of s.e.m. (*P50.05) the anti-apoptotic factor Bcl-2. RT ± PCR analysis was used to examine the e€ect of ADM on the expression of Bcl-2 under normoxia and hypoxia. When added to cultures of Ishikawa cells visualise the induction of Bcl-2 peptide by ADM. A ADM induced a sixfold increase (measured by relative dose dependent up to ®vefold increase of Bcl-2 peptide band density) in the level of Bcl-2 mRNA after 6 h of expression could be observed after hypoxic treatment incubation under hypoxia or normoxia (Figure 7a). To of Ishikawa cells with ADM at concentrations of con®rm these results, Western blots were used to 10712 ±1076 M for 8 h (Figure 7b).

Oncogene Adrenomedullin in tumourogenesis MK Oehler et al. 2940 Bcl-2 increases survival in transiently transfected Ishikawa-cells We ®nally asked whether Bcl-2 has the ability to protect Ishikawa-cells from hypoxic cell death. There- fore Ishikawa cells were transiently transfected with an expression plasmid for Bcl-2. Seventy-two hours after transfection the cells were exposed to hypoxia. Bcl-2 transfected cells showed 25% less death (sub-G1 peak) than controls in ¯ow cytometric evaluation (P50.05) (Figure 9). This observation veri®es the role of Bcl-2 as a suppressor of hypoxic cell death in Ishikawa cells.

Discussion

Our initial experiments with the endometrial carcino- ma cell line Ishikawa clearly demonstrated that ADM belongs to the family of genes which are upregulated under hypoxic stress. We could also show that not only gene transcription but also ADM peptide secretion is increased under hypoxia. In addition, ADM upregulation was mimicked by CoCl2 which is known to change the redox state of the central iron atom of heme similarly to the e€ect of hypoxia (Gleadle et al., 1995). These results con®rm the observations in several animal and cell systems that ADM seems to be a ubiquitously hypoxia- regulated gene (Garayoa et al., 2000; Hofbauer et al., 2000). Hypoxic regulation of ADM is known to be controlled by a common oxygen sensing pathway in which they key player is the HIF-1 transcription factor. HIF-1 is a heterodimer composed of HIF-1a and HIF-1b/ARNT subunits, both representing mem- bers of the PAS (Per, ARNT, Sim) basic ± helix ± loop helix family (Blancher and Harris, 1998). HIF-1 binds to hypoxia response elements (HREs) several of which have been described in the promoter of ADM Figure 4 ADM inhibits hypoxia-induced apoptosis in early and late stages. (a) Annexin V-FITC and PI-staining of un®xed cells (Garayoa et al., 2000). Thus ADM belongs to a followed by FACScan analysis were performed on Ishikawa cells diverse family of HIF-1-regulated genes, including after 36 h of 0.1% O2 with or without ADM. (b) Data represent VEGF, Epo, -like growth factor binding protein the results of the three replicates, with bars indicating s.e.m. 1 (IGFBP-1) and the platelet-derived endothelial cell (*P50.05) growth factor/thymidine phosphorylase (PD-ECGF/ TP) which are all known to be highly expressed in malignant tumours. It is thought that hypoxia functions as a selective pressure for solid tumours, leading to the progression Ishikawa cells stably transfected with ADM to a more malignant phenotype by selecting for an overexpress Bcl-2 apoptosis-resistant population of cells which show a After having shown Bcl-2 expression to be upregulated distinct expression of hypoxia controlled genes by external application of ADM we expected the Bcl-2 (Graeber et al., 1996). As those cells are more expression in Ishikawa cell-clones stably transfected resistant to conventional radio- and chemotherapy with ADM to be increased. Indeed, the cell clones and show high metastatic potential they signi®cantly overexpressing ADM and exhibiting increased resis- impact on clinical response to anticancer therapies and tance to cell death in the hypoxia experiments both prognosis (Hockel et al., 1999; McGill, 1997). overexpressed Bcl-2 up to sixfold in comparison to the However, it is uncertain which hypoxia regulated controls (Figure 8). These results further con®rmed the genes are involved in the development of an increased mechanism of Bcl-2 induction by ADM in Ishikawa malignant phenotype and even less is known about cells. their function.

Oncogene Adrenomedullin in tumourogenesis MK Oehler et al. 2941

Figure 5 Characterization of two Ishikawa cell clones stably transfected with ADM. The clones were shown to overexpress ADM by the combination of (a) immunohistochemistry, (b) FACS analysis and (c) RT ± PCR

Our experiments with Ishikawa cells show for the may also induce vascular growth and promote tumour ®rst time that ADM protects malignant cells from nutrition. hypoxia-induced cell death. Together with our ob- Although our study is the ®rst to describe the anti- servation that Ishikawa cells express the putative apoptotic e€ect of ADM under hypoxia, prior ADM-receptor CRLR ± RAMP2, the production and observations have shown ADM to antagonize serum secretion of ADM in hypoxic areas of tumours deprivation-induced apoptosis in endothelial cells could establish an autocrine/paracrine-mediated me- (Kato et al., 1997). Interestingly, in glomerular chanism rescuing malignant cells from apoptosis. Since mesangial cells ADM was shown to evoke a contrary ADM has additional angiogenic and vasodilator e€ect by increasing apoptosis during serum starvation capacities (Hinson et al., 2000; Zhao et al., 1998), it (Parameswaran et al., 1999). These results suggest a cell

Oncogene Adrenomedullin in tumourogenesis MK Oehler et al. 2942

Figure 8 ADM stably transfected Ishikawa cells overexpress Bcl-2. Total protein was isolated from ADM stably transfected Figure 6 ADM stably transfected Ishikawa cells are more Ishikawa cell clones and a control. Bcl-2 expression was resistant to hypoxic cell death. Annexin V-FITC-staining was quanti®ed by immunoblotting performed on ADM stably transfected Ishikawa clones after 24 h of 0.1% O2. Data represent the results of ®ve independent experiments

Figure 9 Ishikawa cells are more resistant to hypoxic cell death when transiently transfected with Bcl-2. PI staining and quanti®cation of the sub-G1 peak were performed on Ishikawa cells after transient transfection with Bcl-2 and incubation at 0.1% O2 for 24 h. Data represent the mean value of ®ve replicates with bars indicating s.e.m.

Figure 7 ADM upregulates Bcl-2 expression on the mRNA and protein level. (a) Ishikawa cells were exposed to normoxia or hypoxia for 8 h with or without ADM. Total RNA was isolated protective factor against many apoptotic stimuli, and Bcl-2 quanti®ed by RT ± PCR. (b) Bcl-2 including hypoxia (Shimizu et al., 1996). It is thought protein expression was visualized by immunoblotting of cell to block apoptosis by retaining Cytochrome c in the extracts isolated from Ishikawa cells incubated for 8 h at 0.1% O2 mitochondria, thereby inhibiting caspase-3 activation with di€erent ADM concentrations and subsequent downstream apoptotic events (Yang, 1997). In con®rmation, Bcl-2 transfected Ishikawa-cells showed a decreased hypoxic apoptotic rate in our type dependent e€ect of ADM on apoptosis, as has experiments, an e€ect which has been described before been demonstrated for its mitogenic e€ects (Hinson et for other cell types such as hepatoma or endothelium al., 2000). (NoÈ r et al., 1999; Yamabe et al., 1998). We found ADM upregulated the oncoprotein Bcl-2 High Bcl-2-protein levels or aberrant Bcl-2-gene at the mRNA and peptide level in Ishikawa cells under expression have been reported in a wide variety of normoxia and hypoxia. Bcl-2 is a well known solid tumours, including endometrial cancer, and it was

Oncogene Adrenomedullin in tumourogenesis MK Oehler et al. 2943 suggested that disruption of apoptosis as a result of observation that TAM induces hypoxia in xenografts changes of Bcl-2 levels is an important regulatory step combined with our ®ndings indicate an indirect in malignant growth and/or acquisition of more hypoxia-mediated pathway (Evans et al., 1997; Furman aggressive neoplastic phenotypes (Ricca and Biroccio, et al., 1992). 2000). This might explain why Bcl-2 expression is In conclusion, the present study demonstrates that associated with poor prognosis in several types of ADM antagonizes hypoxia induced cell death of malignancies (Sakuragi et al., 1998). endometrial cancer cells by up-regulation of Bcl-2 in Our study is the ®rst to report the anti-apoptotic an autocrine/paracrine manner. These observations, mechanism of Bcl-2 upregulation by ADM. Others taken together with our previous results that ADM is have recently demonstrated other pathways: ADM an angiogenic factor, implicates ADM as a promoter antagonizes serum-deprivation-induced endothelial of tumour growth and a possible target for anticancer apoptosis by upregulation of the transcription factor strategies. Max, the heterodimeric partner of c-Myc (Shichiri et al., 1999). The proto-oncogene c-Myc transcriptionally controls the expression of a diverse group of genes and its deregulation leads to a cellular imbalance in the Materials and methods expression of genes that promote proliferation, trans- formation and apoptosis. For its function as transcrip- Materials and cells tional activator c-Myc is believed to require The human ADM cDNA was a gift from Dr Kitamura, First dimerization with Max. It was anticipated that the Department of Internal Medicine, Kihara, Japan (Kitamura presence of abundant Max ± e.g. upregulated by et al., 1993). The expression plasmid hBcl-2/pcBABE was ADM ± leads to the formation of Max ± Max provided by Dr Norbury, Cell Cycle Group, Institute of homodimers which suppress apoptosis by antagonizing Molecular Medicine, Oxford, UK. Human ADM-(1-52) was the e€ect of the Myc ± Max-transcription complex synthesized by the ICRF-Peptide-Unit (Clare Hall, London, through competition for recognition sites (Shichiri et UK). All reagents were purchased from Sigma (Poole, UK), unless stated otherwise. The steroid receptor positive al., 1999). endometrial carcinoma cell line Ishikawa was provided from Interactions have also been shown between Bcl-2 and the Clare Hall Laboratories (Imperial Cancer Research Fund, Myc. Numerous studies on c-Myc function have shown London, UK). Cells were routinely cultured at 378C and 5% that c-Myc changes the balance between proliferation CO2 in DMEM supplemented with 10% FCS and 4 mM and apoptosis in the direction of programmed cell glutamine. death under conditions of hypoxia or serum depletion (Cory et al., 1999). However, in vivo experiments, Hypoxia and cobalt chloride treatment including bitransgenic mice models, have shown that in Bcl-2 overexpressing tumours the c-Myc induced pro- Hypoxic incubations were performed in an oxygen-regulator apoptotic impulse is counteracted enabling cellular incubator (Forma Scienti®c, Labtech, UK) at 0.1% O2 which re¯ects the oxygen tension in solid tumours (Vaupel, 1993). proliferation (Jager et al., 1997; Pellegrini and Strasser, Cells were grown in DMEM/1% FCS for 24 h before 1999). The exact mechanism by which Bcl-2 interacts incubation at 0.1% O2 in DMEM/1% FCS. with c-Myc is uncertain. It is believed to inhibit Ishikawa cells were treated with 500 mM cobalt chloride for apoptosis downstream of c-Myc since it does not 8 h to mimic hypoxia. appear to a€ect c-Myc mediated p53 protein accumula- tion (Alarcon et al., 1996). Vector construction and stable transfection of Ishikawa cells It is noteworthy that the products of both anti- apoptotic ADM target genes, Bcl-2 and Max, interact The full-length cDNA for human ADM was cloned into with Myc. However whether Bcl-2 and Max are part of pcDNAl/Neo to give plasmid pCMV ± ADMneo in which a common mechanism inhibiting Myc-triggered cell expression of ADM is constitutively driven by the CMV- promoter. Ishikawa cells were transfected with pCMV ± death remains to be determined. ADMneo or pcDNAl/Neo (empty vector control) by ADM expression is upregulated by TAM in human electroporation. Transfected cells were selected for neomycin endometrium (Elkas et al., 2000; Zhao et al., 1998). resistance by treating them with the antibiotic G418 (500 mg/ TAM is used for the treatment of breast cancer and is ml) in DMEM/10% FCS for 12 weeks. Isolated clones being evaluated as a prevention agent in women at were characterized for their ADM-mRNA and peptide high risk of developing the disease (Cuzick, 2000). A expression by RT ± PCR, immunohistochemistry and FACS major concern is the proliferative e€ect of TAM on the analysis. endometrium, which is associated with a 2 ± 7-fold increased risk of endometrial cancer for long-term Transient transfection assays tamoxifen users (Barakat, 1996; Bergman et al., 2000; The full-length cDNA for human Bcl-2 cloned into the Carthew et al., 2000). The mitogenic and antiapoptotic expression plasmid pcBABE and the empty vector were properties of ADM point to this peptide being a transfected into Ishikawa cells using Fugene reagent (Roche, mediator of tamoxifen's adverse endometrial e€ects. Lewes, UK). Transfection eciency was evaluated by The mechanism by which TAM stimulates the ADM cotransfection with green ¯uorescent protein (Invitrogen, gene is uncertain since it lacks palindromic oestrogen Groningen, The Netherlands). Seventy-two hours after response elements (Ishimitsu et al., 1994). The transient transfection Ishikawa cells were exposed to hypoxia.

Oncogene Adrenomedullin in tumourogenesis MK Oehler et al. 2944 (1006) with a Centricon-3-Concentrator followed by a RNA isolation Microcon-5-Concentrator (Amicon, Beverley, USA). Protein Total RNA (derived from 26107 cells) for RNAse protection concentrations were quanti®ed using the BCA protein assay and RT ± PCR experiments was extracted using TRl-reagent (Pierce, Rockford, USA). The samples were stored at 7808C (Sigma) according to the manufacturer's instructions. All until assay. samples were stored at 7708C until assay. Immunoblotting For immunoblotting samples were run on a Synthesis of riboprobes 10% SDS gel for Bcl-2 or on a gradient 10 ± 20% tricine, SDS ± PAGE (BioRad, Hercules, USA) for ADM detection. A 261 bp fragment of the hADM cDNA was subcloned into The proteins were electrophoretically transferred onto a Bluescript Vector and transcribed with T7 RNA polymerase polyvinylidene di¯uoride membrane (Immobilon PVDF ± to create 32P-CTP labelled antisense riboprobes. ADM Millipore, Watford, UK). The PVDF membrane was blocked riboprobes were prepared to the highest speci®c activity for 2 h in PBS with 0.1% Tween-20, 5% bovine serum possible. As loading control U6 small nuclear RNA albumin and incubated in a 1 : 1000 dilution of a monoclonal (GeneBank accession no X01366) was used. To account for mouse anti-human Bcl-2 antibody (DAKO, Glostrup, Den- the signal strength of the highly abundant control a mark) or a polyclonal rabbit anti-human ADM antibody riboprobe of signi®cantly lower speci®city was prepared. (Peninsula Laboratories, Merseyside, UK). Binding was visualized by a horseradish peroxidase-conjugated rabbit RNAse protection assay anti-mouse IgG or pig anti-rabbit IgG, respectively and developed with ECL Western blotting reagents (Amersham, RNAse protection assays followed previously described Little Chalfont, UK). Quanti®cation of protein expression methods (Petersen et al., 1995). Brie¯y, 10 mg total RNA was performed by densitometry (Fluorchem, Alpha Innotech were denatured and hybridized with 32P-CTP labelled ribop- Corp, USA). robes. Single-stranded RNA was digested with RNase H (Promega, Southampton, UK) and the protected fragments size-fractionated on a 6% polyacrylamide gel. The signal Radioimmunoassay intensity was quanti®ed on a phosphoimager (Molecular in the culture medium were extracted according to a Dynamics, Sunnyvale, USA). previously published method using Sep-Pak C18 cartridges (Waters, Milford, USA) (Kato et al., 1997). Human ADM (1- Reverse transcription (RT)-polymerase chain reaction (PCR) 52) was quanti®ed by a competitive RIA (Peninsula Laboratories, Merseyside, UK) using rabbit-antiserum. ADM, CRLR, RAMP2 and Bcl-2 gene expression were Samples or standards were preincubated with antiserum prior examined by RT ± PCR using the Reverse-iT- (®rst strand to addition of tracer. Bound peptide was separated with goat synthesis kit) (Advanced Biotechnologies, Epsom, UK). For anti-rabbit IgG. Assay detection limit was 1 pmol/l. the RT-reaction anchored oligo-dT primers included in the kit were used. cDNA encoding ADM was ampli®ed for 25 cycles (958C Cell death analysis for 60 s, 608C for 30 s and 728C for 60 s). The following The cellular DNA content was evaluated as described by primers were used (GeneBank accession no NM001124): 5'- Ongkeko et al. (1995) with slight modi®cations. Cells were CTGGGTTCGCTCGCCTTCCTA-3' (forward, base 187 to harvested and ®xed in cold 60% ethanol for 1 h at 7208C and 207); 5'-GTTGTCCTTGTCCTTATCTGT-3' (reverse, base then suspended in phosphate bu€ered saline containing 0.1% 558 to 538) resulting in a PCR-fragment of 371 bp. Triton X-100. After the addition of propidium iodide (PI, cDNA encoding CRLR was ampli®ed for 30 cycles (958C 40 mg/ml) and RNAse A (100 mg/ml) the cell suspension was for 60 s, 588C for 30 s and 728C for 90 s). The following incubated for 1 h at 48C in the dark and ®nally analysed using primers were used (GeneBank accession no U17473): 5'- a FACScan (Becton Dickinson, Franklin Lakes, USA). The CTCCTCTACATTATCCATGG-3' (forward, base 1422 to proportion of cells undergoing cell death in each population 1441); 5'-CCTCCTCTGCAATCTTTCC-3' (reverse, base was estimated by quanti®cation of cells with a DNA content 1644 to 1626) resulting in a PCR-fragment of 222 bp. less than that of normoxic cells in G1 (sub-G1 peak). cDNA encoding RAMP2 was ampli®ed for 30 cycles (958C To further distinguish apoptotic from necrotic cells double for 60 s, 568C for 60 s and 728C for 60 s). The following staining for exposed phosphatidylserine and PI exclusion was primers were used (GeneBank accession no NM005854): 5'- performed as follows: Cells were harvested, washed twice ATTGCCTGGAGCACTTTGC-3' (forward, base 361 to with PBS and resuspended in binding bu€er (10 mM HEPES/ 379); 5'-CCTCACTGTCTTTACTCC-3' (reverse, base 587 NaOH, pH 7.4, 140 mM NaCl, 2.5 mM CaCl2). Five ml to 568) resulting in a PCR-fragment of 226 bp. Annexin V-FlITC antibody (Pharmingen, San Diego, USA) cDna encoding Bcl-2 was ampli®ed for 25 cycles (958C for and 10 ml PI (50 mg/ml) were added to the cells. After 60 s, 568C for 30 s and 728C for 90 s). The following primers incubation for 15 min at room temperature in the dark the were used (GeneBank accession no M13994): 5'-TCATT- cells were analysed by a FACScan. The following controls TATCCAGCAGCTTTCGGAAAATG-3' (forward, base 429 were used to set up compensation and quadrants: (a) to 437); 5'-GATTTGAAACTTCCCAATGAATCAGGAGT- unstained cells; (b) cells stained with Annexin V-FITC (no 3' (reverse, base 842 to 814) resulting in a PCR-fragment of PI) and cells stained with PI (no Annexin V-FITC). 413 bp.

Immunoblot analysis Statistical analysis Pepide isolation For isolation of ADM peptide out of cell Statistics were performed using the SPSS software package. culture supernatant, serum-free DMEM was conditioned by Comparisons between groups were made applying the con¯uent cultures for 48 h. The medium was concentrated Wilcoxon matched pairs test.

Oncogene Adrenomedullin in tumourogenesis MK Oehler et al. 2945 Abbreviations Acknowledgments ADM, adrenomedullin; CRLR, calcitonin receptor-like We thank Drs S Tafuro and M Goern (Institute of receptor; Epo, erythropoietin; HIF-1, hypoxia-inducible Molecular Medicine, Oxford) for their technical expertise transcription factor-1; IGFBP-1, insulin-like growth factor andhelpinFACScanexperimentsandHelenTurley binding protein l; PD-ECFG/TP, platelet-derived endothe- (Department of Cellular Science, University of Oxford) lial cell growth factor/thymidine phosphorylase; RAMP2, for help with the immunohistochemistry. Dr MK Oehler receptor associated modifying protein; TAM, tamoxifen; was supported by a postdoctoral fellowship from the VEGF, vascular endothelial growth factor. Deutsche Forschungsgemeinschaft (Oe-230.1-1).

References

Adams JM and Cory S. (1998). Science, 281, 1322 ± 1326. Kitamura K, Kangawa K, Kawamoto M, Ichiki Y, Alarcon RM, Rupnow BA, Graeber TG, Knox SJ and Nakamura S, Matsuo H and Eto T. (1993). Biochem. Giaccia AJ. (1996). Cancer Res., 56, 4315 ± 4319. Biophys. Res. Commun., 192, 553 ± 560. Barakat R. (1996). Clin. Obstet. Gynecol., 39, 629 ± 640. McGill G. (1997). Front Biosci., 2, d353 ± d579. Bergman L, Beelen M, Gallee M, Hollema H, Benraadt J and McLatchie LM, Fraser NJ, Main MJ, Wise A, Brown J, van Leeuwen F. (2000). Lancet, 356, 881 ± 887. Thompson N, Solari R, Lee MG and Foord SM. (1998). Blancher C and Harris AL. (1998). Cancer Metas. Rev., 17, Nature, 393, 333 ± 339. 187 ± 194. Miller MJ, Martinez A, Unsworth EJ, Thiele CJ, Moody Caron KM and Smithies O. (2001). Proc. Natl. Acad. Sci. TW, Elsasser T and Cuttitta F. (1996). J. Biol. Chem, 271, USA. 98, 615 ± 619. 23345 ± 23351. Carthew P, Edwards R, Nolan B, Martin E, Heydon R, Naik P, Karrim J and Hanahan D. (1996). Genes Dev., 10, White and Tucker M. (2000). Carcinogenesis, 21, 793 ± 2105. 797. NakayamaM,TakahashiK,MurakamiO,ShiratoKand Cormier-Regard S, Nguyen SV and Claycomb WC. (1998). Shibahara S. (1998). Biochem. Biophys. Res. Commun., J. Biol. Chem., 273, 17787 ± 17792. 243, 514 ± 517. Cory S, Vaux DL, Strasser A, Harris AW and Adams JM. Nguyen SV and Claycomb WC. (1999). Biochem. Biophys. (1999). Cancer Res., 59, 1685s ± 1692s. Res. Commun., 265, 382 ± 386. Cuzick J. (2000). Eur. J. Cancer, 36, 1298 ± 1302. NoÈ r JE, Christensen J, Mooney DJ and Polverini PJ. (1999). Elkas J, Armstrong A, Pohl J, Cuttitta F, Martinez A and Am. J. Pathol., 154, 375 ± 384. Gray K. (2000). Obstet. Gynecol., 95, 697 ± 703. Ongkeko W, Ferguson DJ, Harris AL and Norbury C. Evans SM, Koch CJ, Laughlin KM, Jenkins WT, Van (1995). J. Cell. Sci., 108, 2897 ± 2904. Winkle T and Wilson DF. (1997). Cancer Res., 57, 5155 ± Parameswaran N, Nambi P, Brooks DP and Spielman WS. 5161. (1999). Eur. J. Pharmacol., 372, 85 ± 95. Furman E, Rushkin E, Margalit R, Bendel P and Degani H. Pellegrini M and Strasser A. (1999). J. Clin. Immunol., 19, (1992). J. Steroid Biochem. Mol. Biol., 43, 189 ± 195. 365 ± 377. Garayoa M, Martinez A, Lee S, Pio R, An WG, Neckers L, Petersen N, Larsen L, Nissen H, Jensen L, Jensen A, Petersen Trepel J, Montuenga LM, Ryan H, Johnson R, Gassmann P, Horder M, Gregersen N and Kristiansen K. (1995). M and Cuttitta F. (2000). Mol. Endocrinol., 14, 848 ± 862. Clin. Chem., 41, 1605 ± 1613. Gleadle JM, Ebert BL, Firth JD and Ratcli€e PJ. (1995). Am. Ricca A and Biroccio A. (2000). Int. J. Cancer, 86, 188 ± 196. J. Physiol., 268, C1362 ± C1368. Sakuragi N, Ohkouchi T and Hareyama H. (1998). Int. J. Graeber TG, Osmanian C, Jacks T, Housman DE, Koch CJ, Cancer, 79, 153 ± 158. Lowe SW and Giaccia AJ. (1996). Nature, 379, 88 ± 91. Shichiri M, Kato H, Doi M, Marumo F and Hirata Y. Hague S, Zhang L, Oehler MK, Manek S, MacKenzie IZ, (1999). Mol. Endocrinol., 13, 1353 ± 1363. Bicknell R and Rees MC. (2000). Clin. Cancer Res., 6, ShimizuS,EguchiY,KamiikeW,ItohY,HasegawaJ, 2808 ± 2814. Yamabe K, Otsuki Y, Matsuda H and Tsujimoto Y. Hinson JP, Kapas S and Smith DM. (2000). Endocr. Rev., 21, (1996). Cancer Res., 56, 2161 ± 2166. 138 ± 167. ShimizuS,EguchiY,KosakaH,KamiikeW,MatsudaH Hockel M, Schlenger K, Hockel S and Vaupel P. (1999). and Tsujimoto Y. (1995). Nature, 374, 811 ± 813. Cancer Res., 59, 4525 ± 4528. Thomlinson R and Gray L. (1995). Br.J.Cancer,9, 539 ± Hofbauer KH, Jensen BL, Kurtz A and Sandner P. (2000). 549. Am. J. Physiol. Regul. Integr. Comp. Physiol., 278, R513 ± Vaupel P. (1993). Drug Resistance in Oncology. Teicher B R519. (ed.). Marcel Dekker: New York, pp. 53 ± 85. Ishimitsu T, Kojima M, Kangawa K, Hino J, Matsuoka H, Wimalawnsa SJ. (1997). Crit. Rev. Neurobiol., 11, 167 ± 239. Kitamura K, Eto T and Matsuo H. (1994). Biochem. Yamabe K. Shimizu S, Kamiike W, Waguri S, Eguchi Y, Biophys. Res. Commun., 203, 631 ± 539. Hasegawa J, Okuno S, Yoshioka Y, Ito T, Sawa Y, Jager R, Herzer U, Schenkel J and Weiher H. (1997). Uchiyama Y, Tsujimoto Y and Matsuda H. (1998). Oncogene, 15, 1787 ± 1795. Biochem. Biophys. Res. Commun., 243, 217 ± 223. Kamitani S, Asakawa M, Shimekake Y, Kuwasako K, Yang J. (1997). Science, 275, 1129 ± 1132. Nakahara K and Sakata T. (1999). FEBS Lett., 448, 111 ± Zhao Y, Hague S, Manek S, Zhang L, Bicknell R and Rees 114. MC. (1998). Oncogene, 16, 409 ± 415. Kato H, Shichiri M, Marumo F and Hirata Y. (1997). Endocrinology, 138, 2615 ± 2620.

Oncogene