Orexins Acting at Native OX 1 Receptor in Colon Cancer and Neuroblastoma Cells or at Recombinant OX 1 Receptor Suppress Cell Growth by Inducing Apoptosis*

Screening of 26 gut peptides for their ability to inhibit growth of human cancer HT29-D4 cells grown in 10% fetal calf serum identified orexin-A and orexin-B as anti-growth factors. Upon addition of either orexin (1 (cid:1) M ), suppression of cell growth was total after 24 h and > 70% after 48 or 72 h, with an EC 50 of 5 n M peptide. Orexins did not alter proliferation but promoted apoptosis as demonstrated by morphological changes in cell shape, DNA fragmentation, chromatin condensation, cytochrome c release into cytosol, and activation of caspase-3 and caspase-7. The serpentine G protein-cou-pled orexin receptor OX 1 R but not OX 2 R was expressed in HT29-D4 cells and mediated orexin-induced Ca 2 (cid:2) transients in HT29-D4 cells. The expression of OX 1 induction of cell apoptosis. Flow cytometric data of cell cycle analysis of control cells cultured in standard medium with 10% FCS indicated that 69.5 (cid:6) 4.5% of HT29-D4 cells are in G 0 /G 1 phase; 15.6 (cid:6) 2.9% are in S phase, and 13.2 (cid:6) 2.2% are in G 2 /M phase ( n (cid:3) 6). When HT29-D4 cells were treated with 1 (cid:2) M orexin-A or orexin-B for 24 h, 69.0 (cid:6) 5.3 and 70.2 (cid:6) 5.9% of cells were found in G 0 /G 1 phase, 14.8 (cid:6) 2.2 and 13.4 (cid:6) 2.1% in S phase, and 12.3 (cid:6) 3.6 and 14.0 (cid:6) 2.5% in G 2 /M phase, respectively ( n (cid:3) 6). These data indicated that orexins have no significant effect on the HT29-D4 cell cycle. This was confirmed when experiments were carried out with HT29-D4 cells that were synchronized by serum deprivation for 48 h. in S, and 4.6 (cid:6) 0.3% in G 2 /M) (with FCS: 57.5 (cid:6) 3.3% in G 0 /G 1 , 23.4 (cid:6) 1.3% in S, and 16.3 (cid:6) 0.9% in G 2 /M) (with FCS and orexin-B: 56.9 (cid:6) 2.9% in G 0 /G 1 , 25.1 (cid:6) 2.1% in S ,and 16.2 (cid:6) 0.6% in G 2 /M)). B shows effect of orexin-B (1 (cid:2) M ) on [ methyl - 3 H]thymidine incorporation in the DNA of HT29-D4 cells grown in the presence of 10% FCS. Cells were seeded at 10 5 cells per well in 12-well plates and grown for 24 h in standard culture medium with 10% FCS. Cells were then grown in serum-free medium for 72 h. Finally, cells were cultured for up to 24 h in fresh medium containing 10% FCS and 0.1 (cid:2) Ci of [ methyl - 3 H]thymidine in the presence ( black bar ) or in the absence ( open bar ) of 1 (cid:2) M orexin-B. At time 0, cells were 427,000 (cid:6) 28,000 per well. After 24 h, HT29-D4 cells were 801,000 (cid:6) 33,900 in standard medium and 408,000 (cid:6) 10,680 in the presence orexin-B. disintegrations of [ methyl - 3 H]thymidine incorporated triplicate

Classical growth factors for colon cancer cells have been extensively described including agonists of tyrosine kinase receptors such as epidermal growth factor and related proteins (1) or insulin-like growth factors (2). More recently, some G protein-coupled receptor (GPCR) 1 agonists such as peptide hormones (3)(4)(5), prostaglandins (6), or serine proteases (7,8) have been shown also to promote colon cancer cell proliferation often through transactivation of the epidermal growth factor receptor (6,8). These GPCRs are expressed in both normal colonic epithelium and colon tumors (9) or even ectopically expressed by cancer cells such as in the case of the neurotensin receptor NT1 (10) or the thrombin receptor protease-activated receptor 1 (7). Whatever their expression pattern, they probably all contribute to the growth of colon tumors because of the presence of abundant ligands in the neuroendocrine environment of colonic tumors and/or to the production of receptor ligands by the tumor itself (11,12).
Our knowledge of receptor agonist suppressing colon cancer cell growth is much more limited apart from a few observations regarding transforming growth factor-␤ (13) or Fas ligand (14). We reasoned that among the very rich environment of peptide hormones and neuropeptides in the gut, we should be able to find natural agonists behaving as suppressors of colon cancer growth. In order to test this hypothesis, we developed a very simple assay by using human colon adenocarcinoma cells HT29-D4 grown in 10% FCS and screened for various peptides by their ability to inhibit cell growth. We made two dramatic hits with orexin-A and orexin-B, which appear to be robust growth inhibitors as shown here.
Orexin-A and orexin-B (15), also named hypocretin-1 and hypocretin-2 (16), were discovered in 1998 by orphan receptor technologies (15) or subtractive cDNA cloning (16). They are encoded by a single gene that drives the synthesis of preproorexin that is subsequently matured into the 33-amino acid orexin-A and the 28-amino acid orexin-B, sharing 46% amino acid identity in humans (reviewed in Ref. 17). Two orexin receptor subtypes OX 1 R and OX 2 R have been cloned (15). They are serpentine GPCRs that bind both orexins with poor selectivity and are coupled to Ca 2ϩ mobilization (15). Orexins were initially characterized as neuropeptides restricted to hypothalamic neurons that project in the brain to nuclei involved in the control of feeding, sleep-awakeness, neuroendocrine homeostasis, and autonomic regulation (17). Genetic or experimental alterations of the orexin system have been shown to be associated with narcolepsy (18,19). More recent observations indicated that orexins and their receptors are not restricted to the hypothalamus but are also expressed in a few peripheral tissues (17), including the gastrointestinal tract (20).
Here we show that orexins acting at the OX 1 R suppress the growth of human colon cancer cells HT29-D4 by promoting apoptosis through cytochrome c release from mitochondria and caspase activation. We further expand upon these data by showing that activation of native OX 1 R in other colon cancer cell lines and neuroblastoma cells or activation of recombinant OX 1 R expressed in CHO cells also leads to strong apoptosis and subsequent growth suppression. The OX 1 receptor-mediated apoptosis therefore appears to be an intrinsic property of the receptor regardless of the cell type where the receptor is expressed.
Isolation of Colonic Epithelial Cells from Normal Human Colon-Fresh normal human colons with no digestive disease were collected with the assistance of France-Transplant following French bioethical law. The colons were removed from small intestine and then immediately carried from the operating room to our laboratory in an isothermic box on ice. It usually took 30 -60 min from the colon collection to the beginning of the epithelial cell isolation procedure. The colons were gently washed with water, and normal colon epithelial cell isolation was performed as reported previously (23).
Explant Culture of Human Colonic Mucosa-Fragments of human normal sigmoid colon taken at about 10 cm downstream to the tumor were obtained from a patient undergoing surgery for moderately differentiated colon carcinoma. The tissue fragments were processed according to the Guidelines of the French Ethics Committee for Research on Human Tissues. A sample, taken adjacent to the explants, was submit-ted to histological analysis and subsequently reported as normal by the pathologists. Immediately after removal, the tissues were placed in 4°C oxygenated Krebs solution and processed as described previously (24). The muscularis propria was removed by microdissection under microscopic control in a Sylgard-coated Petri dish. Mucosa was then carefully stripped from the underlying compartment made of muscularis mucosae and submucosa. Fragments of 20 -30 mg were cut out and pinned in Sylgard-coated Petri dishes and maintained in culture for 24 h in 2 ml of RPMI 1640 medium (Invitrogen) containing 0.01% bovine serum albumin and antibiotics (200 g/ml streptomycin, 200 units/ml penicillin, 1% fungizone, Invitrogen). The explants were maintained at 37°C in a 95% oxygen, 5% carbon dioxide humid atmosphere on a rocking platform at 30 rpm, in the absence (n ϭ 4) or presence (n ϭ 4) of 1 M orexin-B. Before culture (t 0 ), some dissected fragments were frozen for subsequent analysis. At the end of the 24-h culture, the supernatants were centrifuged, and aliquots were stored at Ϫ80°C for further analysis. Tissue specimens were cut into several fragments as follows: one was used for intracellular ATP measurement; one was used for histological examination after formalin fixation/paraffin-embedding, and one was stored at Ϫ80°C. Tissue integrity and cell viability of the explants were assessed by standard morphological analysis, i.e. by hematoxylin-eosin staining on paraffin sections. Cell viability was also assessed by two assays as follows: measurement of the percentage of LDH released and of the intracellular ATP level. LDH was measured with the Enzyline LDH kit both in the supernatant after the 24-h incubation (extracellular LDH or LDHe) and in a small fragment of the tissue (intracellular LDH or LDHi). Percent toxicity, represented by percentage LDH released, was calculated as (LDHe/(LDHe ϩ LDHi) ϫ 100), taking into account the weight of the tissue samples. The LDH release was not significantly different upon a 24-h orexin-B treatment from control cultures (8.5 Ϯ 1.2 versus 10.5 Ϯ 1% respectively, n ϭ 4). Cell viability was also assessed in the mucosal explants by measuring the intracellular ATP level before (t 0 ) and after the 24-h culture. ATP was measured in tissue lysates with a luminometric assay ATPlite kit according to the manufacturer's instructions. Orexin-B (1 M) did not significantly modify the ATP level (7 Ϯ 0.9 pmol/g protein, n ϭ 4) compared with the untreated explant cultures (5.5 Ϯ 0.6 pmol/g protein, n ϭ 4). These levels were not significantly different from those of the explanted tissue at t 0 , i.e. before culture (6.2 Ϯ 0.7 pmol/g protein).
Cell Growth Assay-CHO-K cells, CHO/hOX 1 R cells (both seeded at 5 ϫ 10 4 cells/well), HT29-D4 cells, Caco-2, SW480, LoVo, HCT116, and SK-N-MC cells (seeded at 2 ϫ 10 5 cells/well) were grown in 24-well plates for 24 h in standard culture conditions with 10% FCS (see above). The culture medium was then replaced every 24 h with fresh medium containing or not containing orexins at concentrations indicated in the legends to the figures. At the end of the treatment, adherent cells were trypsinized, and cells excluding trypan blue were counted in a hemocytometer.
Cell Cycle Analysis and in Vitro [methyl-3 H]Thymidine Incorporation Assay-HT29-D4 cells (3 ϫ 10 5 ) were seeded in 25-cm 2 dishes and grown at 70 -80% confluency in standard medium. In some experiments, cells were maintained in serum-free media for 48 h in order to synchronize the cell cycle. The culture medium was then replaced with fresh standard medium containing or not containing 1 M orexins. After 24 h, adherent cells were harvested by trypsinization and fixed at Ϫ20°C by 70% ethanol in phosphate-buffered saline (PBS). Cell aliquots (10 6 cells) were then treated with 50 g/ml RNase A for 15 min and stained with 10 g/ml propidium iodide. Percentages of cells in G 0 /G 1 , S, and G 2 /M phases were determined by flow cytofluorometric analysis (Beckman Coulter Epics XL-MCL). For [methyl-3 H]thymidine incorporation, HT29-D4 cells (10 5 cells/well) were seeded in 12-well clusters (Costar) in medium with 10% FCS and cultured for 24 h to allow cell adhesion. The medium was then removed; cells were washed with PBS and cultured for another 48 h with serum-free medium. Thereafter, orexin-B (1 M) was added in standard (10% FCS) or serumfree medium, and cells were cultured for 1, 2, 4, 6, or 24 h in the presence of 0.1 Ci of [methyl-3 H]thymidine per well. The medium was then removed, and cells were washed twice with PBS and incubated in trypsin/EDTA for 10 min. They were then harvested in 200 l of medium, incubated for 30 min in 5% trichloroacetate, and centrifuged at 10,000 ϫ g for 4 min. Pellets were washed in 95% ethanol, solubilized in 10% Triton X-100, and put in 2.5 ml of scintillation fluid for counting incorporated radioactivity. For each point, cells excluding trypan blue were counted in a hemocytometer. The experiments were performed in triplicate wells and at least repeated twice. Results are expressed in disintegrations/min/10 6 viable cells.
Characterization of Apoptosis in Cultured Cell Lines-Three methods were used for the characterization of apoptosis in cultured cell lines.
For the TUNEL method, cells (7 ϫ 10 5 ) were seeded on Lab-Tek chamber coverglasses and grown for 24 h in standard culture medium. The culture medium was then replaced with fresh culture medium containing or not containing orexins. After 24 h, cells were fixed for 30 min in 4% paraformaldehyde/PBS at 4°C followed by permeabilization for 10 min with 0.2% Triton X-100 in PBS. Cells were then analyzed by the TUNEL method using the Roche Diagnostics in situ cell death detection kit according to the manufacturer's instructions. The samples were vizualized with a confocal laser scanning microscope (LSM 510 META, Zeiss). For in situ chromosome analysis, cells (2 ϫ 10 5 ) were seeded on glass coverslips and grown for 24 h in standard medium. The culture medium was then replaced with fresh culture medium containing or not containing orexins. After 24 h, cells were fixed with 80% ethanol for 10 min at 4°C and stained with either propidium iodide (50 g/ml) or DAPI (10 g/ml). Apoptotic nuclei were detected by epifluorescence microscopy. For DNA fragmentation assay, cells (2 ϫ 10 6 ) were seeded in culture dishes and grown for 24 h in standard culture medium. The culture medium was then replaced with fresh culture medium containing or not containing orexins. After 48 h, cells were trypsinized and collected by centrifugation for 5 min at 1,000 ϫ g. DNA was then extracted as described (25). Briefly, cells were washed in 1 ml of cold PBS and then lysed in 50 mM Tris-HCl, pH 8.0, containing 2 mM NaCl, 10 mM EDTA, 4% SDS, and 0.38 mg/ml of proteinase K. Samples were incubated for 30 min at 65°C, followed by an overnight incubation at 37°C. DNA was then precipitated, washed, resuspended in Tris-EDTA buffer, and incubated for 1 h with 0.1 mg/ml RNase A as described (25). DNA fragmentation was analyzed on 1.5% agarose gels in the presence of 0.5 mg/ml ethidium bromide.
Cytochrome c Release-HT29-D4 cells were seeded in 75-cm 2 culture dishes (10 6 /dish) and grown in standard culture medium for 24 h. The medium was then replaced by a fresh culture medium containing or not containing orexins, and cells were further grown for 24 h. After cell lysis, cytochrome c was measured in cytosol (14,000 ϫ g supernatant after 15 min centrifugation) by using an ELISA kit (see above) according to the manufacturer's instructions. Under these experimental conditions, the cytosolic fraction did not contain mitochondria as verified by electronic microscopy. 2 Alternatively, cytochrome c was characterized by immunoprecipitation followed by Western blot. Briefly, cytochrome c in cytosol was immunoprecipitated using the 6H2-B4 monoclonal anticytochrome c antibodies (2 g/ml) and protein G-Sepharose beads. After addition of SDS-PAGE sample buffer to the washed beads, the samples were separated by electrophoresis on 10% SDS-polyacrylamide gel and then transferred to nitrocellulose membrane. Total cytochrome c present in the cellular extract before cytosolic separation was characterized as described above. The blot was probed with 1 g/ml of 7H8 -2C12 monoclonal anti-cytochrome c antibodies, and immune complexes were revealed with secondary peroxidase-conjugated antibodies using a chemiluminescent kit.
In Situ Caspase Activation-HT29-D4 cells (7 ϫ 10 5 ) were seeded on Lab-Tek chamber coverglasses and grown for 24 h in standard culture medium. The culture medium was then replaced with fresh culture medium containing or not containing 1 M orexins. After 24 h, caspase activation was detected as described (26) using the CaspACE TM FITC-VAD-FMK In Situ Marker (Promega Corp., Madison, WI) which binds to activated caspases. The bound marker was then localized by fluorescence detection using a confocal microscope.
Immunocytochemical Studies of Cleaved Forms of Caspase-3 and Caspase-7-HT29-D4 cells (7 ϫ 10 5 ) were seeded on Lab-Tek chamber cover glasses and grown for 24 h in standard culture medium. The culture medium was then replaced with fresh culture medium containing or not containing orexins. After 24 h, the coverslips were probed with rabbit polyclonal anti-cleaved caspase-3 antibodies (Asp-175, dilution 1:100) or anti-cleaved caspase-7 antibodies (Asp-198, dilution 1:100) that specifically recognize cleaved enzyme isoforms. These antibodies do not recognize full-length caspase-3 or full-length caspase-7 or other cleaved caspases. Fluorescein isothiocyanate (FITC)-conjugated goat anti-rabbit immunoglobulin IgG was used as the secondary antibody. Vectashield mounting medium containing propidium iodide was added, and coverslips were observed by confocal microscopy.
Western Immunoblotting Studies of Caspase-3 and Caspase-7-HT29-D4 cells were seeded in 75-cm 2 culture dishes (10 6 /dish) and grown in standard culture medium for 24 h. The culture medium was then replaced with fresh culture medium containing or not containing orexins. After 24 h, cells were lysed by adding Chaps cell extract buffer containing 50 mM Pipes/NaOH (pH 6.5), 2 mM EDTA, 0.1% Chaps, 5 mM DTT, 2 g/ml leupeptin, 1 g/ml pepstatin, 1 g/ml aprotinin, and 1 mM phenylmethylsulfonyl fluoride. Cells were resuspended in the buffer and frozen and thawed three times, and the lysate was centrifuged for 30 min at 14,000 rpm. After addition of SDS-PAGE sample buffer to the supernatant (30 g of protein), the samples were separated by electrophoresis on 16% SDS-polyacrylamide gel and then transferred to nitrocellulose membrane. Blots were probed with rabbit polyclonal anticleaved caspase-3 antibodies (Asp-175, dilution 1:1000) or anti-cleaved caspase-7 antibodies (Asp-198, dilution 1:1000) that specifically recognize cleaved enzyme isoforms. Subsequently, blots were probed with rabbit polyclonal anti-caspase-3 antibodies (1:1000), which principally detect the full-length procaspase-3 (32 kDa) and the fragment of cleaved caspase-3 following cleavage at Asp-175 (17 kDa), or with rabbit polyclonal anti-caspase-7 antibodies (1:1000), which detect the full-length procaspase-7 (35 kDa). Blots were standardized by using the mouse monoclonal anti-␤-actin antibodies (AC-74). Immune complexes were revealed with secondary peroxidase-conjugated antibodies by using a chemiluminescent kit.
Morphological Analysis of Apoptotic Cell Death in Human Colonic Mucosa Explants-Morphological analysis of apoptotic cell death was assessed on paraffin sections by using two assays. The DNA-specific dye Hoechst 33258 (Calbiochem), which visualizes all the steps of the apoptotic process, was applied on deparaffinized sections (1 g/ml in Hanks' balanced salt solution without phenol red, Invitrogen). Sections were mounted with the Prolong Antifade medium (Molecular Probes, Eugene, OR). The fluorescence was observed on an Axiovert 200-M-Carl Zeiss microscope equipped with an ApoTome slider. Cells were visualized with a ϫ63/1.4 oil immersion lens. Image processing was performed using an AxioCam MRCCD camera and the AxioVision 4.0 software (Carl Zeiss). In addition, immunohistochemical staining of apoptotic epithelial cells was performed by using the M30 antibody (1:50) as described previously (27). This antibody recognizes a cytokeratin 18 neoepitope cleaved by caspases and is considered as an early marker of apoptosis in epithelial cells (28,29).
Apoptotic Cell Measurement by Annexin V Labeling-Apoptosis was analyzed in SK-N-MC, CHO/hOX 1 R, and parent CHO-K cell lines using the Guava Nexin TM kit, which discriminates between apoptotic and nonapoptotic cells. The Guava Nexin TM assay utilizes annexin V-phycoerythrin to detect phosphatidylserines on the external membrane of apoptotic cells. Cells (2 ϫ 10 6 ) were seeded in culture dishes and grown for 24 h in standard culture medium. The culture medium was then replaced with fresh culture medium containing or not containing orexin-B. After 24 h, apoptotic cell staining was performed according to the manufacturer's instructions and analyzed with a Guava PCA system. Results are expressed as the percentage of apoptotic annexin V-phycoerythrin-positive cells and are the means of four analyses.
Immunocytochemical Detection of OX 1 R-Cells (2 ϫ 10 5 ) were cultured on coverslips, washed with cold PBS, and immediately fixed in ice-cold 4% paraformaldehyde in PBS for 1 h followed by three washings in PBS. Samples were incubated for 1 h at 4°C with 10% newborn calf serum in TBST buffer (10 mM Tris, pH 8, 150 mM NaCl and 0.1% Tween 20) and then incubated overnight at 4°C with rabbit polyclonal antibodies against OX 1 R (dilution 1:100). Staining was revealed using FITC-conjugated goat serum anti-rabbit IgG by confocal microscopy.
Fura-2/AM Loading and Intracellular Calcium Measurement by Fluorimeter-Intracellular calcium concentrations were measured using Fura-2/AM. HT29-D4 cells or HCT116 cells (5 ϫ 10 3 cells/ml) were seeded onto the center of glass coverslips and cultured in DMEM for 4 days to 70 -80% confluence. These coverslips were then loaded with 5 M of Fura-2/AM in Na-Hepes-buffered saline, pH 7.4 (135 mM NaCl, 4.6 mM KCl, 1.2 mM MgCl 2 , 11 mM Hepes, 11 mM glucose, and 1.5 mM CaCl 2 ), containing 0.01% pluronic acid for 45-60 min at 37°C. They were then washed in Na-Hepes buffer and placed at 37°C in a fluorimeter. The fluorescence was measured with a dual wavelength excitation fluorimeter at 340 and 380 nm for excitation and 510 nm for emission. The cells were challenged first with 1 M orexin-B followed by a control challenge with 0.1 M neurotensin.
Miscellaneous Procedures-Routine procedures such as radioimmunoassay of intracellular cAMP content (30) and measurement of protein contents (31) were performed as described.
Statistical Analysis-All data were expressed as mean Ϯ S.E. values and analyzed by analysis of variance and Student's t test for statistical significance. A p value of Ͻ0.05 was considered as statistically significant.

Orexins Inhibit Cell Growth in Human Colon Cancer Cells
HT29-D4 in Culture-Human colon cancer HT29-D4 cells grown in standard medium in the presence of 10% FCS were treated for 24 h with a variety of peptide hormones or neuropeptides present in the gut. Among the 26 peptides tested, orexin-A and orexin-B were the only peptides inhibiting cell growth, other peptides being without any effect or even stimulating cell growth such as in the case of ghrelin (Table I). In the presence of 10% FCS, which triggers a strong mitogenic effect on HT29-D4 cells, orexin-A and orexin-B elicited a dramatic decrease in cell number (Fig. 1, A and B). This inhibition of serum-induced increase in cell number, referred to as suppression of cell growth, was almost total after 24 h (Fig. 1, A and B) of treatment and still extensive after 48 (Fig. 1, A and B) or 72 h (Fig. 1B) of treatment, i.e. Ͼ70% suppression. Orexins were active in the range of concentrations between 1 nM and 1 M, with half-maximal responses being obtained for 5 nM for both orexins (Fig. 1C). Similar dose-response curves were observed after challenging cells with orexins for 24 ( Fig. 1C) or 48 h (not shown). After 1-2 days of challenge with orexins some morphological changes in cell shape were observed, in particular the cells looked rounder and shrunken (Fig. 1A, insets f and g). Although no cell detachment was observed, these morphological changes were reminiscent of apoptosis.
Orexins Promote Cell Apoptosis but Do Not Alter Cell Proliferation in Human Colon Cancer Cells HT29-D4 in Culture-Next we determined whether suppression of HT29-D4 cell growth by orexins was related to inhibition of cell proliferation and/or induction of cell apoptosis. Flow cytometric data of cell cycle analysis of control cells cultured in standard medium with 10% FCS indicated that 69.5 Ϯ 4.5% of HT29-D4 cells are in G 0 /G 1 phase; 15.6 Ϯ 2.9% are in S phase, and 13.2 Ϯ 2.2% are in G 2 /M phase (n ϭ 6). When HT29-D4 cells were treated with 1 M orexin-A or orexin-B for 24 h, 69.0 Ϯ 5.3 and 70.2 Ϯ 5.9% of cells were found in G 0 /G 1 phase, 14.8 Ϯ 2.2 and 13.4 Ϯ 2.1% in S phase, and 12.3 Ϯ 3.6 and 14.0 Ϯ 2.5% in G 2 /M phase, respectively (n ϭ 6). These data indicated that orexins have no significant effect on the HT29-D4 cell cycle. This was further confirmed when experiments were carried out with HT29-D4 cells that were synchronized by serum deprivation for 48 h. Indeed treatment of synchronized cells with 10% serum in the absence or presence of 1 M orexin-B ( Fig. 2A) or orexin-A (data not shown) provided identical flow cytometric data. Direct assessment of DNA synthesis by [methyl-3 H]thymidine incorporation into DNA of HT29-D4 cells further supported the idea that orexins did not alter cell proliferation. Indeed, treatment of synchronized HT29-D4 cells with orexin-B (1 M) did not modify serum-induced [methyl-3 H]thymidine incorporation into DNA (Fig. 2B).
In sharp contrast with the proliferation data, a body of evidence supported that HT29-D4 cells undergo apoptosis upon orexin-A or orexin-B treatment. Fluorescence microscopic analysis of DNA-staining patterns with DAPI (Fig. 3A, a-c) revealed an increase in apoptotic cell death in orexin-treated cells as compared with control cells. Moreover, HT29-D4 cells displayed typical chromatin condensation and fragmentation of nuclei into small spherical particles upon treatment with orexins. DNA fragmentation was ascertained by the TUNEL assay (Fig. 3A, d-f). TUNEL-positive cells were observed upon orexin-A or orexin-B treatment, whereas no labeling was detected in control cells. Apoptosis was further indicated by typical DNA ladder corresponding to cleavage of genomic DNA upon cell treatment with orexins (Fig. 3B).
Orexin-induced Cell Apoptosis in Human Colon Cancer HT29-D4 Cells Is Associated with Cytochrome c Release and Caspase Activation-Orexin-induced apoptosis was shown to be associated with cytochrome c release into cytosol. By using a cytochrome c ELISA, we showed that cytochrome c levels in cytosol were significantly increased 3, 15, and 24 h after challenging the cells with orexin-A or orexin-B (Fig. 4A). These data were consistent with immunoprecipitation/Western blotting analysis that also indicated a sharp increase in cytosolic cyto- chrome c after a 24-h challenge with orexins (Fig. 4B). Further experiments showed that orexin-induced apoptosis is associated with caspase activation. In situ caspase activation was followed by cleavage of a fluorogenic substrate (see "Experimental Procedures"). The fluorescent product was strongly labeled orexin-B (1 M)-treated cells (Fig. 5A, b). Similar results were obtained with orexin-A (data not shown). A faint fluorescent labeling was also observed in control cells indicating a low background of caspase activation in the culture conditions used (Fig. 5A, a). To confirm caspase activation in orexin-induced apoptosis in HT29-D4 cells, we immunodetected cleavage of effector caspases downstream of the cytochrome c release, i.e. caspase-3 and caspase-7. By using cleaved caspase-3 (Asp-175) and cleaved caspase-7 (Asp-198) antibodies, endogenous levels of cleaved caspase-3 (Fig. 5A, d) and cleaved caspase-7 (Fig. 5A, f) were detected in HT29-D4 apoptotic cells upon 24 h of treatment with 1 M orexin-B. Cleaved caspase-3 and caspase-7 detected upon HT29-D4 cell treatment with orexin-B were colocalized with fragmented nuclei (Fig. 5A, d and f). Similar results were obtained with orexin-A (data not shown). No cleaved caspase-3 (Fig. 5A, c) or caspase-7 (Fig. 5A, e) could be detected in control untreated cells. Further characterization of the effects of orexins on caspase cleavage was obtained by Western blot. As shown in Fig. 5B, treatment of HT29-D4 cells for 24 h with 1 M orexin-B resulted in the appearance of 19and 17-kDa forms of cleaved caspase-3 and the 20-kDa form of cleaved caspase-7 (Fig. 5B, right), whereas no cleaved caspases could be detected in control cells. Similar data were obtained with orexin-A (not shown). The pro-forms of caspase-3 (Fig. 5B, left) and caspase-7 (Fig. 5B, right) were present in both control and orexin-treated cells. HT29-D4 Cells Express the OX 1 Receptor Subtype-Because reliable orexin tracers are still unavailable (17), the nature of orexin receptor subtype(s) expressed in HT29-D4 cells was determined by RT-PCR and immunocytochemistry. Total mRNA from HT29-D4 cells or control CHO cell lines expressing recombinant hOX 1 R or hOX 2 R were reverse-transcribed and amplified with specific couples of primers for the two subtypes of orexin receptors. Amplification products of the expected size were obtained in HT29-D4 cells with OX 1 R primers but not OX 2 R primers (Fig. 6A). Control recombinant CHO cells clearly express OX 1 R or OX 2 R transcripts (Fig. 6A), supporting the idea that the absence of OX 2 R mRNA in HT29-D4 cells was not related to limits of RT-PCR technology. The OX 1 R-amplified product obtained from HT29-D4 cells was sequenced and found to be identical to the hOX 1 R cDNA sequence (GenBank TM AF0412343). To provide evidence for OX 1 R protein expression in HT29-D4 cells, indirect immunofluorescence was performed by using rabbit polyclonal antibodies against hOX 1 R (Fig. 6B). Strong immunostaining with membrane localization was observed. This staining was completely abolished by co-incubating antibodies with the immunogen peptide.
Because orexins were shown previously to consistently induce Ca 2ϩ transients in orexin receptor-expressing cells (15), we further investigated this second messenger as evidence for the presence of functional OX 1 R in HT29-D4 cells. The intracellular Ca 2ϩ concentration was first monitored under a confocal microscope using Fluo-4AM dye. Fig. 6C shows confocal pictures before and after treatment with 1 M orexin-B. The neuropeptide clearly induced Ca 2ϩ transients as shown by confocal images. The effect of orexin-B on cytosolic calcium was further tested in Fura-2/AM dye-loaded HT-29-D4 cells by fluorescence analysis in a classical fluorimeter. Again orexin-B induced calcium transients with a time course of response similar to that observed with neurotensin (Fig. 6D), a well known inducer of calcium transients in human colon cancer cells (32). In sharp contrast, orexin-B failed to induce calcium transients in the human colon cancer HCT116 cells (Fig. 6D) that are not equipped with OX 1 receptor (see below). As a control, we showed that neurotensin nicely induced calcium transients in HCT116 cells. All these results clearly showed that functional OX 1 receptors are expressed in HT29-D4 cells. Fig. 7, the OX 1 R-mediated effects of orexins are observed in three of four other human colon cancer cell lines tested. RT-PCR experiments showed that amplification products of the expected size were obtained in Caco-2, SW480, and LoVo cells with OX 1 R primers (Fig. 7A). In contrast, no OX 1 R mRNA could be detected in the HCT116 cell line. In good agreement with the RT-PCR data, orexin-B strongly inhibited FCS-stimulated cell growth in Caco-2, SW480, and LoVo cells but not in HCT116 cells (Fig. 7B). Finally, the effect of orexin-B on apoptosis was tested in the four colon cancer cell lines. Apoptosis was clearly indicated by a typical DNA ladder corresponding to cleavage of genomic DNA upon cell treatment with orexin-B in Caco-2, SW480, and LoVo but not HCT116 cells (Fig. 7C). Altogether these data indicated that expression of OX 1 R and OX 1 R-mediated anti-growth and pro-apoptotic effects of orexins are frequent in colon cancer because they are observed in four of five human colon cancer cell lines originating from different patient tumors, i.e. HT29-D4, Caco-2, SW480, and LoVo.

Expression of OX 1 R and Anti-growth and Pro-apoptotic Effects of Orexins Are Observed in Other Colon Cancer Cell Lines but Not in Normal Colonic Epithelium-As shown in
In this context, we explored the status of OX 1 R in normal human colonic mucosa. RT-PCR experiments using total RNA extracted from pure epithelial cell preparations isolated from three normal human colons failed to indicate OX 1 R mRNA (not shown) under conditions in which specific amplification products were clearly detected in human colon cancer cell lines (see Figs. 6A and 7A). Because the long term culture of isolated human colonic epithelial cells still remains an elusive task and the existence of normal intestinal epithelial cell lines is still a debated question, we used explant cultures of dissected human normal colonic mucosa to assess the effects of orexin on apoptosis in normal colon. These mucosal explants (polarized epithelial barrier and underlying lamina propria) can be maintained in good viability conditions, as assessed by a multiparametric approach (see "Experimental Procedures"). Mucosal explants maintained their morphological integrity over a 24-h culture period as shown by standard histology (Fig.  8). Spontaneous apoptosis occurred in a minority of epithelial cells (less than 1%). Indeed, a few apoptotic epithelial cells were visualized by the M30 antibody immunostaining specific for caspase-3-cleaved cytokeratin 18 and by the Hoechst dye, which reveals DNA condensation and fragmentation. They were preferentially located at the tip of the surface epithelium, undergoing exfoliation, and occasionally in the lower region of the crypt (Fig. 8). A few lamina propria macrophages underlying the epithelial barrier occasionally contained M30 antibodypositive cells. Most interestingly, a 24-h treatment with 1 M orexin-B neither altered the morphological integrity of the colonic crypts nor increased the number of apoptotic cells (Fig. 8).
The OX 1 R-mediated Anti-growth and Pro-apoptotic Effects of Orexins Are Intrinsic Properties of the OX 1 R and Not Dependent on Cell Context-All the actions of orexins described herein were observed in human colon cancer cell lines. We next asked the question of whether the OX 1 R-mediated anti-growth and pro-apoptotic effects of orexins are restricted to colon cancer cells or are intrinsic properties of the OX 1 R. Because OX 1 Rs were initially described in the brain (15), we first tested the human neuroblastoma cell line SK-N-MC which expresses OX 1 R (33). As shown in Fig. 9, orexin-A and orexin-B, in the range of concentrations between 1 nM and 1 M, strongly reduced SK-N-MC cell growth (Fig. 9A, top). The maximal effect observed at 1 M peptides represented 75% inhibition of growth as compared with control cells. Half-maximal inhibitions were obtained for 5 nM for both orexins (Fig. 9A, top). The neuropeptides also induced SK-N-MC cell apoptosis as shown by propidium iodide staining, which revealed condensed nuclei upon orexin challenge but not in control cells (Fig. 9B).
To generalize further the role of OX 1 R in inhibition of cell growth and induction of apoptosis, we considered CHO cells expressing the recombinant hOX 1 R. Quite interestingly, orexin-A and orexin-B strongly inhibited CHO/hOX 1 R cell growth (Fig. 9A, middle), whereas no effect of orexins on cell growth could be detected in the parent CHO-K cell line (Fig.  9A, bottom). Similarly, propidium iodide staining indicated that both orexins triggered apoptosis in the CHO/hOX 1 R cell line (Fig. 9B), whereas parent CHO-K cells did not undergo apoptosis upon orexin treatment (Fig. 9B).
In order to better quantitate the apoptotic rate of SK-N-MC and CHO/hOX 1 R cells upon orexin challenge, we analyzed annexin V binding (Guava nexin assay), which reveals phosphatidylserine externalization in apoptotic cells. After a 24-h treatment of cells with 1 M orexin-B, the percentage of apoptotic cells strongly increased up to 11 and 27% in SK-N-MC and CHO/hOX 1 R cells, respectively (Table II). In sharp contrast, orexin-B had no significant effect on the percentage of apoptotic cells in the parent CHO-K cell line (Table II). Altogether these results supported the idea that OX 1 R mediates inhibition of cell growth and induction of apoptosis independently of the cell environment.

DISCUSSION
In this work, we discover and characterize a new function of the neuropeptides orexins as drastic pro-apoptotic peptides. We show that orexins acting at either native (human colon adenocarcinoma cells or neuroblastoma cells) or recombinant (CHO cells) seven transmembrane domain receptor OX 1 R have strong anti-growth properties by inducing cytochrome c-and caspasedependent apoptosis. These data provide the first evidence that the OX 1 R and its natural agonists orexin-A and orexin-B are important players in the control of apoptosis. They may have future interesting applications in the treatment of apoptosisresistant cancers such as colon cancer. They also promote the design of new studies to understand the physiological role of orexins in relation to apoptosis in OX 1 R-expressing tissues. Considering that many growth factors (1-8) but very few anti-growth factors for human colon cancer cells have been described during the last 2 decades (13, 14, 34), we tested a large array of gut peptide hormones and neuropeptides for their ability to inhibit cell growth in the human colon cancer HT29-D4 cell line cultured in the strong growth-promoting environment of 10% FCS. All peptides tested except orexins were without any effect in inhibiting cell growth. Among the peptides tested, some were described previously as growthpromoting such as neurotensin (3), glucagon-like peptide 1 (35), or gastrin/cholecystokinin (5). It is not surprising that they are inactive in our assay conditions because growth is already strongly stimulated by FCS. In this context, the ability of ghrelin, a peptide hormone produced by the stomach (36), to promote cell growth even in the presence of 10% FCS is amazing but is outside the scope of the present study. The screening assay identified only two anti-growth peptides, the closely related orexin-A and orexin-B encoded by the same gene (15,17) and acting at common receptors (15). Initial experiments established that orexin-A and orexin-B are equipotent at suppressing HT29-D4 cell growth (see Fig. 1) with an ED 50  reminiscent of apoptosis, are observed, in particular cells look rounder and shrunken. (iii) Direct evidence for orexin-induced apoptosis is provided by several techniques including DNAstaining patterns with DAPI, TUNEL assay, and DNA ladder. (iv) Orexins induce cytochrome c release into cytosol and activation of effector caspase-3 and caspase-7. The orexin-induced apoptosis in HT29-D4 cells is not associated with the observable cell detachment in contrast to Fas ligand-induced apoptosis (37), which is responsible for massive HT29-D4 cell de-tachment. 3 Because cells do undergo apoptosis without detachment, we have to assume that apoptotic cells in culture are eliminated. Although not documented in this paper, a possible way is phagocytosis by adjacent nonapoptotic epithelial cells, a process demonstrated previously (38,39).
Two human orexin receptors sharing 64% amino acid identity have been cloned (15). They are serpentine GPCRs coupled to Ca 2ϩ mobilization (15,40). Because their molecular pharmacology is poorly developed (17) and reliable orexin tracers are still unavailable, orexin receptors have been mainly characterized in tissues or cell lines because of their coupling to calcium mobilization (15,40), by studies of receptor mRNA expression (20,41), or by immunocytochemical detection of receptor proteins (20). In this context, all the currently available tools indicate that HT29-D4 cells are only equipped with OX 1 R. (i) RT-PCR experiments identified an OX 1 R transcript whose sequence is identical to that of OX 1 R cDNA. In sharp contrast, no OX 2 R transcript is detected in HT29-D4 cells under RT-PCR conditions that nicely provide a specific product in control OX 2 R-expressing CHO cells. (ii) the OX 1 R protein is immunodetected at the cell surface of HT29-D4 cells by using a specific anti-OX 1 R antibody (20). In contrast, specific anti-OX 2 R antibodies do not immunostain HT29-D4 cells under conditions in which control OX 2 R-expressing CHO cells are nicely stained. 3 The presence of a functional orexin receptor in HT29-D4 cells is otherwise indicated by robust calcium response observed upon treatment with orexins. The presence of only OX 1 R in HT29-D4 cells strongly suggests that this receptor mediates the pro-apoptotic effects of orexins. This OX 1 R does not discriminate between orexin-A and orexin-B because the two peptides are equipotent at suppressing HT29-D4 cell growth (see Fig. 1). The pharmacology of orexin receptors is still in its early stages, and the literature is somewhat inconsistent regarding the ability of OX 1 R to discriminate between the two orexins. In calcium mobilization studies using CHO cells expressing recombinant OX 1 R, orexin-A was shown to be from 5-to 100-fold more potent than orexin-B (15,(42)(43)(44) contributing to the disparity of their reported selective actions. In the same transfected cellular model, binding studies resulted in contradictory results. By using 125 I-labeled orexin-A as a tracer, the affinity of orexin-A was found to be 20 times higher than orexin-B (15). In contrast, Wieland et al. (33) found identical affinity for the two orexins in CHO cells expressing OX 1 R when using 125 I-labeled orexin-A as a tracer. More unexpectedly, orexin-B was shown to have even higher affinity than orexin-A when using 125 I-labeled orexin-B as a tracer in CHO cells expressing recombinant OX 1 R or SK-N-MC neuroblastoma cells expressing native OX 1 R (33). In this context, further studies and the development of instrumental agonists and/or antagonists of OX 1 R are clearly needed to ascertain the pharmacological profile of OX 1 R.
The role of OX 1 R in controlling apoptosis is directly demonstrated by the expression of recombinant OX 1 R in CHO-K cells that confers the ability of orexins to promote apoptosis in this cell line. This experiment also suggests that the proapoptotic role of OX 1 R is an intrinsic property of the receptor that is not restricted to the HT29-D4 cell context. In line with this idea, our work shows that the OX 1 R-expressing SK-N-MC neuroblastoma cells (33) do undergo apoptosis upon treatment with orexins.
The pro-apoptotic activity of orexins demonstrated here in human colon cancer cells HT29-D4, human neuroblastoma cells SK-N-MC and CHO cells expressing recombinant OX 1 R receptor, results in massive suppression of cell growth regardless of the OX 1 R-expressing cell. The mechanisms of coupling between activation of cell surface seven transmembrane domain OX 1 R and release of cytochrome c into cytosol remain to be elucidated. Activation of OX 1 R is known to result in mobilization of intracellular calcium through a G q -dependent mechanism (15,40,45). This Ca 2ϩ response also observed here in HT29-D4 cells is certainly not sufficient to explain the proapoptotic effect of orexins, although increases of cytosolic calcium are well known to occur during apoptotic cell death (46). Indeed, a variety of GPCRs in human colon cancer cells including HT-29 cells is known to promote intracellular Ca 2ϩ mobilization (3,7,8,10). These receptors such as NT1 receptors for neurotensin (3,10), protease-activated receptors 1 (PAR1) for thrombin (7), protease-activated receptors 2 (PAR2) for trypsin (8), or muscarinic M3 receptors for acetylcholine (47) not only do not trigger apoptosis but rather stimulate cell proliferation. In this context, it is worth pointing out that up to now very few serpentine GPCRs have been shown to inhibit cell growth by promoting apoptosis e.g. endothelin receptor ET B (48), chemo-

TABLE II Detection of apoptotic cell by annexin V labeling in SK-N-MC,
CHO/hOX 1 R, and parent CHO-K cell lines Apoptosis was analyzed in cell lines using the Guava Nexin™ kit, which discriminates between apoptotic and nonapoptotic cells. The Guava Nexin™ assay utilizes annexin V-phycoerythrin (annexin V-PE) to detect phosphatydylserine on the external membrane of apoptotic cells. Cells (2 ϫ 10 6 ) were seeded in culture dishes and grown for 24 h in standard culture medium. The culture medium was then replaced with fresh culture medium containing or not orexin-B (1 M). After 24 h, apoptotic cell staining was performed according to the manufacturer's instructions by using a Guava PCA system. Results are expressed as the percentage of apoptotic annexin V-PEby positive cells and are the means of four analyses. kine CXCR4 receptor (49), ␤-adrenergic receptor (50), angiotensin AT2 receptor (51), somatostatin SST2 receptor (52), or parathyroid hormone receptor (53).
The OX 1 R-mediated pro-apoptotic role of orexins described herein raises the question of its significance in health and disease. With respect to colon cancer, this study already shows that expression of OX 1 R and OX 1 R-mediated apoptosis are frequent in colon cancer cells because they are observed in four of five human colon cancer cell lines originating from different patients (9), i.e. HT29-D4, Caco-2, SW480, and LoVo. Most interestingly, our data also show that normal human colonic epithelial cells are not equipped with OX 1 R, resulting in the absence of the pro-apoptotic effect of orexins in normal human colon epithelium. These observations suggest that OX 1 Rs are aberrantly expressed in human colon cancer cells. Whether this aberrant expression represents an ectopic expression or the re-expression of receptors expressed in colonic epithelium during embryogenesis remains to be determined. Ectopic expression of G protein-coupled receptors in colon cancers has been reported previously for neurotensin receptor (10) and PAR1 for thrombin (7). However, these receptors were clearly shown to promote cell proliferation and colon cancer cell growth (7,10). In that respect, this paper reports the first example of an ectopic expression of a receptor promoting apoptosis in colon cancer cells. Given the known resistance of colon cancer to apoptosis and chemotherapy (39,54), OX 1 R thereby represents an attractive new target for the development of instrumental orexin agonists in this cancer. On the other hand, because the pro-apoptotic role of OX 1 R appears to be an intrinsic property of the receptor regardless of the nature of the OX 1 R-expressing cell, we may speculate about the role of orexins in normal tissues expressing OX 1 R. In that respect, a site of expression of orexin receptors in normal tissues is small intestinal epithelium (17,20). Because this epithelium is a rapidly renewing tissue in which cell homeostasis is regulated by a balance among proliferation, growth arrest, differentiation, and apoptosis (39), the possibility that orexins, which are neuropeptides expressed in the small intestinal wall (20), may control apoptosis and cell homeostasis in this tissue is an attractive hypothesis. A major site of expression of OX 1 R is the brain and more specifically the hypothalamus (15,16). In this context, our studies raise the question of the possible role of orexins in neuronal apoptosis which is a major event during brain development and maturation (55) and also in neurodegenerative diseases (55).
In conclusion, this work characterizes for the first time the pro-apoptotic and subsequent anti-cell growth properties of orexins acting at the OX 1 receptor. Although the molecular mechanisms of the OX 1 R-mediated pro-apoptotic effect of orexins remain to be elucidated, these findings add a new dimension to the biological activities of these neuropeptides that may have important implications in health and disease.