| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
INSERM, U773, Centre de Recherche Biomédicale Bichat Beaujon CRB3, and Université Paris Diderot, UMR S 773, Paris, France
1Correspondence: INSERM, U773, CRB3, Paris, F-75018, France; Université Paris Diderot, UMR S 773, F-75018, Paris, France. E-mail: tvoisin{at}bichat.inserm.fr
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
q subunit and transfected with OX1R cDNA, whereas they promoted apoptosis in cells equipped with Gq and OX1R; and 4) the phospholipase C inhibitor U-73122 blocked orexin-stimulated inositol phosphate formation, whereas it had no effect on orexin-induced apoptosis in CHO cells expressing OX1R. These data unravel a novel mechanism, whereby ITIM-expressing GPCRs may trigger apoptosis.—Voisin, T., El Firar, A., Rouyer-Fessard, C., Gratio, V., Laburthe, M. A hallmark of immunoreceptor, the tyrosine-based inhibitory motif ITIM, is present in the G protein-coupled receptor OX1R for orexins and drives apoptosis: a novel mechanism.
Key Words: hypocretin • cancer cell death • colon cancer • tyrosine phosphatase • SHP-2
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, also named hypocretin-1 and hypocretin-2 (2)
, are derived from a common prepropeptide (see ref. 3
for review). In the brain, orexins are produced by hypothalamic neurons that project to and excite many central areas (4)
, in particular, those governing arousal and wakefulness (5)
. The importance of orexins in the control of sleep is highlighted by the deficit in orexigenic function in animal and human narcolepsy (5)
. More recent studies showed that orexins are not restricted to the hypothalamus but are also expressed in peripheral tissues, including adrenals, gastrointestinal tract, and endocrine pancreas (3)
.
An unexpected and fascinating aspect of orexins has been recently highlighted in cancer cells (6)
. Orexins induce dramatic apoptosis in human colon cancer cell lines, resulting in massive reduction of cell growth (6)
. The effect was also observed in human neuroblastoma cells (6)
and rat pancreatic tumor cells (7)
. Orexin-stimulated apoptosis is associated with mitochondrial cytochrome c release into cytosol and activation of caspase-3 and caspase-7 (6)
and appears to be mediated by the OX1 receptor in colon cancer cells and neuroblastoma cells (6)
and by the OX2 receptor in pancreatic tumor cells (7)
. Promotion of apoptosis by orexins is an intrinsic property of orexin receptors since transfection of OX1 or OX2 receptor cDNAs in Chinese hamster ovary (CHO) cells devoid of endogenous receptors is sufficient to confer the ability of orexins to promote apoptosis (6
, 7)
. These recent findings add a new dimension to the biological role of orexins.
Two orexin receptor subtypes OX1R and OX2R have been cloned (1
, 2)
. They are serpentine G protein-coupled receptors (GPCRs) that recognize both orexins with poor selectivity (1)
. In the current absence of reliable binding assay (3)
, very few information is available regarding orexin receptors. What is known is that activation of OX1R or OX2R induces calcium transients in CHO cells expressing recombinant receptors (1
, 6)
. The origin of calcium on orexin challenge has been documented for OX1R. Two calcium pathways have been identified: a receptor-operated calcium influx, possibly mediated by TRPC channels (8)
and an InsP3-mediated [Ca2+]i transient from intracellular pools (1
, 6
, 9)
. In consonance with the regulation of cytoplasmic calcium levels by orexins, it is thought that OX1 receptors are coupled to Gq heterotrimeric proteins (1)
, but it has been also suggested that the OX1 receptor is linked to influx of Ca2+ through a signal pathway independent of Gq protein activation (10)
.
Although cellular calcium overload or perturbation of intracellular Ca2+ compartmentalization can participate in the onset of apoptosis (11)
, the Ca2+ response to orexins is certainly not sufficient to explain the OX1R-mediated apoptotic effect of orexins in human colon cancer cells (6)
. Indeed, a variety of GPCRs in colon cancer cells, such as NTS1 receptor for neurotensin (12)
, protease-activated receptors for thrombin or trypsin (13)
or muscarinic M3 receptor for acetylcholine (14)
, are known to promote the increase of intracellular calcium. These receptors not only do not trigger apoptosis but rather stimulate cell proliferation. In this context, we speculated that the apoptotic action of orexin, mediated by the OX1 receptor, might be driven by an original mechanism not yet identified in GPCRs. Since promotion of apoptosis by orexins appears to be an intrinsic property of the OX1 receptor (see above), and few GPCRs mediate agonist-induced apoptosis (15
16
17)
, we analyzed the sequence of the OX1 receptor for identification of new motifs, which might be associated with its ability to trigger apoptosis. This search resulted in the identification of a canonical immunoreceptor tyrosine-based inhibition motif (ITIM) in the intracellular domain connecting the seventh transmembrane helix to the C-terminal tail of the OX1 receptor. The consensus 6-amino acid ITIM sequence (Ile/Val/Leu/Ser)-X-Tyr-X-X-(Ile/Leu/Val), where X denotes any amino acid, is not considered to be a signature of GPCRs but represents a hallmark of immune inhibitory receptors on lymphoid and myeloid cells, the immunoglobulin G Fc-receptor Fc
RII being prototypical of such receptors (18)
. Studies on immune receptors have defined a general paradigm for ITIM function, in which ligand engagement by inhibitory receptors results in ITIM phosphorylation by Src-like kinases and recruitment of phosphotyrosine or inositol-phosphatases, which are activated thereby (19)
.
Here, we show that the ITIM in the OX1R plays a crucial role in the proapoptotic action of orexins by recruiting the tyrosine phosphatase SHP-2. We also document the permissive role of the Gq protein in the OX1R-mediated apoptotic response, which is independent of the conventional activation of phospholipase C. This represents a novel mechanism, whereby activation of a GPCR results in apoptosis.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Plasmid construction, mutagenesis, and stable expression in CHO cells
A 1.3-kb EcoRI/EcoRI cDNA fragment of the human wild-type OX1R (hOX1R WT) was cloned into the expression vector pEYFP with the yellow fluorescent protein expressed in the C-terminal position. A mutant OX1R Y358F was constructed by PCR using the QuikChange site-directed mutagenesis kit (Stratagene, La Jolla, CA, USA). Mutation was verified by sequencing. CHO clone S cells [referred as to CHO cells, mouse embryonic fibroblastic (MEF) cells, or MEF Q11 cells] were transfected with the different constructs as described previously (6
, 7)
. After 48 h, yellow fluorescent protein (YFP) fluorescence was used to check receptor expression, and transfected CHO cells were selected in the presence of geneticin (G418) at a final concentration of 1 mg/ml for 6 wk and cloned.
Cell culture
CHO cells were grown as previously described (6
, 7)
. MEF cells and MEF Q11 cells were grown as previously described (20)
. The human colon cancer cell line HT29 was obtained from American Type Culture Collection (Rockville, MD, USA) and was grown as described (6)
. All cells were maintained at 37°C in a humidified 5% CO2/air incubator.
Cell growth assay and characterization of apoptosis
CHO cells (seeded at 5x104 cells/well) and HT29 cells (seeded at 2x105 cells/well) were grown in 24-well plates for 24 h in standard culture conditions with FCS (6)
. The culture medium was then replaced every 24 h with fresh medium with or without orexins at concentrations indicated in the figures legends. At the end of the treatment (48 h), adherent cells were trypsinized, and cells excluding trypan blue were counted. Apoptosis was determined using the Guava Nexin kit, which discriminates between apoptotic and nonapoptotic cells (6
, 7)
. Cell staining was performed according to the manufacturer’s instructions and analyzed with a Guava PCA system, 2000 events being counted. Results are expressed as the percentage of apoptotic Annexin-V-phycoerythrin (PE) -positive cells.
Preparation of particulate fraction and binding of 125I-orexin-A to cell membrane
Cells were grown to confluence, washed 3x with 0.13 M PBS (pH 7), harvested and centrifuged at 2000 g for 5 min at 4°C. The cell pellet was then exposed for 30 min at 4°C to hypoosmotic 5 mM HEPES buffer (pH 7.4) containing protease inhibitors. This particulate fraction is referred to as the membrane preparation. 125I-orexin-A (74 TBq/mmol) was prepared with 125I-Na (IMS300, GE Heathcare, Les Ulis, France), according to the chloramine T method and purified on a Sephadex G-50 column. For the binding assay, membranes (200 µg protein/ml) were incubated for 30 min at 20°C in 250 µl of 20 mM TES binding buffer (pH 7.4), containing 0.5% (w/v) BSA, 5 mM KCl, 1 mM CaCl2, 1.2 mM MgCl2, 0.44 mM KH2PO4, 4.2 mM NaHCO3, 10 mM glucose, 1 mM probenecid, and 0.001% (v/v) Tween 20, in the presence of 0.25 nM 125I-orexin-A with or without unlabeled orexin-A. Bound and free peptides were separated by centrifugation. The nonspecific binding, measured in the presence of 1 µM unlabeled orexin-A, represented 
5% of total binding.
Assessment of cell surface expression of receptors
Cell surface expression of receptors was assessed using the rabbit polyclonal antibodies against OX1R (dilution 1: 200). CHO cells grown in 24-well trays were rinsed 2x with 50 mM Tris-HCl (pH 7.7, 100 mM NaCl, 5 mM KCl, 2 mM CaCl2, 5% heat-inactivated horse serum, and 0.5% heat-inactivated fetal bovine serum as described (21)
, and then incubated with anti-OX1R antibodies for 4 h in TES binding buffer. Cells were then washed 3x with the above-described buffer and exposed to the radiolabeled (400,000 cpm/well) second antibodies (125I-labeled goat anti-rabbit IgG). Cells were rinsed again 4x and then lysed with 1M NaOH, and the radioactivity of the lysate was counted. Nonspecific binding was determined with cells that were incubated only with the 125I-labeled second antibody. This method has been described previously for assessment of cell surface expression of GPCRs (21
, 22)
.
Inositol phosphate (InsP) formation assay
Subconfluent cells were labeled with myo3H-inositol (3.15 TBq/mmol) (TRK883, GE Heathcare, Les Ulis, France) for 24 h in DMEM. Cells were then incubated at 37°C in TES binding buffer (pH 7.4) containing 20 mM lithium chloride to block inositol monophosphatase activity. Where indicated, orexin-A (1 µM) was added 1 min before lithium addition and incubation was continued for 30 min. Cells were extracted with ice-cold formic acid and total InsPs separated from free myo3H-inositol using column chromatography.
Immunoprecipitation of OX1R
Assessment of receptor phosphorylation
CHO cells were pretreated 24 h with 50 µM SHP1/2 inhibitor NSC-87877 in the absence or the presence of 10 µM Src-like enzyme tyrosine kinase inhibitor PP2. Semiconfluent cells were then treated with 1 µM orexin-A in fresh culture medium at 37°C for 5 min. Cells were collected and lysed in 50 mM Tris-HCl buffer pH 7.4 containing 0.25% Na deoxycholate, 150 mM NaCl, 1% Nonidet P-40 and 1 mM EGTA. Proteins (500 µg) were then incubated with 2 µg of anti-OX1R antibodies overnight at 4°C. Protein immunoprecipitation was performed according to the manufacturer’s instructions using the Seize X protein G immunoprecipitation kit (Pierce, Rockford, IL, USA). Then, immunoprecipitated proteins suspended in Laemmli buffer were loaded onto a 10% SDS-polyacrylamide gel, transferred to a nitrocellulose membrane, and immunoblotted with anti-OX1R (1:1000 dilution) or antiphosphotyrosine (1:1000 dilution) antibodies. Immune complexes were revealed with secondary peroxidase-conjugated antibodies, using a chemiluminescent kit.
Interaction with SHP-2
Semiconfluent cells were treated with 1 µM orexin-A in fresh culture medium at 37°C for 5 min. Cells were collected and lysed in 50 mM Tris-HCl buffer pH 7.4 containing 0.25% Na deoxycholate, 150 mM NaCl, 1% Nonidet P-40, and 1 mM EGTA. Proteins (500 µg) were then incubated with 2 µg of anti-OX1R antibodies overnight at 4°C. Protein immunoprecipitation was performed according to the manufacturer’s instructions using the Seize X protein G immunoprecipitation kit (Pierce). Then, immunoprecipitated proteins suspended in Laemmli buffer were loaded onto a 10% SDS-polyacrylamide gel, transferred to a nitrocellulose membrane, and immunoblotted with anti-OX1R (1:1000 dilution), anti-SHP-1 (1:1000 dilution), or anti-SHP-2 (1:1000 dilution) antibodies. Immune complexes were revealed with secondary peroxidase-conjugated antibodies, using a chemiluminescent kit.
Miscellaneous
Intracellular calcium concentrations were measured using Fura-2/AM as previously described (6)
. Human C453S mutant SHP-1 or human C459S mutant SHP-2 (23)
were transiently transfected in CHO/hOX1R WT cells expressing the human wild-type OX1 receptor subtype or HT-29 cells as previously described (6
, 7)
.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Fig. 7
) was mutated into phenylalanine. The mutated cDNA was stably transfected in CHO cells and, after cell cloning, a CHO/hOX1R Y358F cell line was obtained. The mutated and wild type receptor constructs being fused in the C-terminal position with YFP, it could be verified that they exhibited similar expression by fluorescence microscopy (not shown). Measurement of annexin V binding to apoptotic cells showed that the Y358F receptor mutant was unable to promote apoptosis whatever the orexin-A concentration up to 10 µM, whereas a typical dose-response curve of orexin-A was observed for the wild-type receptor with an ED50 of 9 ± 3 nM (Fig. 1
A). We have previously shown that orexins have no effect on cell cycle and proliferation but inhibit cell growth through their ability to promote apoptosis (6)
. Orexin-A tested up to 10 µM did not alter the growth of CHO/hOX1R Y358F cells, whereas a typical dose-inhibition curve was obtained in CHO/hOX1R WT cells with an IC50 of 23 ± 5 nM (Fig. 1B
).
|
|
|
|
|
|
|
Though YFP fluorescence (not shown) indicated that the mutated and wild-type receptors were similarly expressed by CHO cells, it was quite important to demonstrate that the Y358F receptor mutant, which no longer mediated orexin-induced apoptosis was expressed at the plasma membrane and still bound orexins. In a first set of experiments, we measured the binding of rabbit polyclonal anti-OX1R antibody to intact CHO cells expressing the wild-type and mutated receptor (Fig. 1C
). As revealed with a secondary rabbit 125I-anti-IgG, the binding of the anti-OX1R antibody was similar on intact CHO cells expressing the wild-type or mutated receptor, whereas very low nonspecific binding was observed on intact mock-transfected cells, which did not express OX1 receptors (Fig. 1C
). In a second set of experiments, we developed a reliable binding assay for OX1 receptors using 125I-orexin-A as a tracer. Scatchard analysis showed that orexin-A bound to a single population of saturable binding sites on CHO/hOX1R WT cells (Fig. 1D
) with a receptor density (Bmax) of 1.4 ± 0.1 pmol/mg of protein and a dissociation constant (Kd) of 6 ± 2 nM. Orexin also bound to a single class of binding sites on CHO/hOX1R Y358F cells (Fig. 1D
). The Bmax (1.3±0.1 pmol/mg of protein) was the same as that measured on CHO cells expressing the WT receptor. However, the Kd of 42 ± 7 nM was 7-fold higher than that of the WT receptor. These data showed that the inability of the Y358F OX1 receptor to mediate orexin-induced apoptosis was neither due to an absence of expression of the mutated receptor on cell surface nor to an alteration of the binding capacity of CHO cells expressing this mutant receptor. Since the mutation abolished the apoptotic response regardless of the orexin concentration up to 10 µM, it is also unlikely that the lower affinity of the mutated receptor for orexin as compared to the wild-type receptor, is responsible for the absence of the apoptotic response. However, the decrease of affinity for orexin of the receptor mutated in its ITIM raised the question of the relationship of ITIM with classical transduction of OX1 receptor through interaction with the Gq protein inasmuch as the interaction of GPCRs with G proteins is crucial for determining the affinity state of GPCRs (24)
.
Role of Gq protein and Gq signaling in recombinant OX1 receptor-mediated apoptosis
The OX1 receptor is coupled to Gq leading to an increase of intracellular calcium transients on orexin challenge (1
, 6
, 25
, 26)
. As measured by fluorescence analysis of Fura-2/AM dye-loaded cells, 1 µM orexin-A (Fig. 2
A) or orexin-B (not shown) induced calcium transients in CHO/hOX1R WT cells but not in CHO/hOX1R Y358F cells, suggesting that the receptor mutant is uncoupled from Gq proteins. This was further evidenced by the inability of 1 µM orexin-A (Fig. 2B
) or orexin-B (not shown) to stimulate inositol phosphate production in CHO/hOX1R Y358F cells, whereas a 10-fold stimulation was observed in CHO/hOX1R WT cells. More direct assessment of the interaction of G protein with receptors was obtained by measuring the ability of the GTP analog Gpp(NH)p to inhibit 125I-orexin-A binding (Fig. 2C
). Gpp(NH)p considerably reduced 125I-orexin-A binding to the wild-type receptor but was without any effect on tracer binding to the receptor mutant. Interestingly, Scatchard analysis (not shown) indicated that the affinity state of the WT receptor in the presence of Gpp(NH)p was similar to that of the mutant receptor; i.e., Kd of 53 ± 9 and 42 ± 7 nM, respectively. This suggested that mutation of tyrosine in the ITIM uncouples the OX1 receptor from G protein resulting in a low-affinity state of the receptor similar to that triggered by Gpp(NH)p.
The ITIM of OX1R is located in the intracellular domain connecting the seventh transmembrane helix to the C-terminal tail (Fig. 7)
, a domain that is known to be important for interaction of Gq with GPCRs (27)
. Whether the ITIM participates directly in the recruitment of Gq or simply overlaps a site of interaction of OX1R with Gq is unclear. In this context, we decided to develop an alternative strategy to understand the role of Gq in the OX1R-mediated apoptosis. The OX1R was transfected in either MEF cells or cells in which the Gq subunit was invalidated (MEF Q11 cells). As shown in Fig. 3
A, orexin-A promoted apoptosis in MEF cells but not in MEF Q11 cells transfected with the OX1R. These data supported a role of Gq in OX1R-mediated apoptosis. Does it follow that classical Gq signaling through activation of phospholipase C is involved in orexin-induced apoptosis? Going back to the CHO/hOX1R cell model, we tested this hypothesis by using the specific inhibitor of phospholipase C, U-73122. We found that U-73122 up to 10 µM did not inhibit apoptosis triggered by orexin-A, whereas it dose dependently inhibited inositol phosphate formation triggered by orexin-A, the inhibition being complete at 10 µM inhibitor (Fig. 3B
). U-73122 (10 µM) also reversed orexin-A (1 µM) -induced calcium transient in CHO/hOX1R cells (not shown). These data demonstrated that orexin-induced apoptosis is dependent on Gq but not on the conventional Gq downstream pathway resulting in phospholiase C activation, inositol phosphate formation and induction of calcium transients.
Orexin-induced apoptosis involves a protein tyrosine phosphatase in the recombinant CHO OX1R-expressing cells
Since ITIMs recruit phosphatases in immune receptors (19)
, we evaluated the impact of the protein tyrosine phosphatase inhibitor phenyl arsine oxide (PAO) on orexin-induced apoptosis in CHO cells expressing the recombinant WT human OX1R. After 48-h treatment of cells grown in 5% FCS with orexin-A (1 µM), an important apoptosis was revealed by annexin-V binding to externalized phosphatidylserine (Fig. 4
A). Quantitation indicated that 28.1 ± 0.4% of cells were apoptotic on orexin-A treatment vs. 3.0 ± 0.4% in control cells. In the presence of 1 µM PAO, the proapoptotic action of orexin-A was completely reversed (Fig. 4A
), with a proportion of apoptotic cells of 2.5 ± 0.6%. Very similar data were obtained when orexin-B (1 µM) was used as an agonist of OX1R (not shown), with a proportion of apoptotic cells of 26.3 ± 0.5 and 3.2 ± 0.5% in the absence and presence of 1 µM PAO, respectively. We also tested the effect of PAO on the orexin-mediated inhibition of CHO/hOX1R WT cell growth (Fig. 4B
). A 48-h culture in the presence of 5% FCS results in notable cell growth. Orexin-A (1 µM) induced an extensive inhibition of cell growth, which was completely reversed in the presence of 1 µM PAO (Fig. 4B
). PAO alone had no effect on cell growth (not shown). Similar data were obtained when orexin-B (1 µM) was used as a OX1R agonist (not shown). Altogether, these data suggest that a phosphotyrosine protein phosphatase is involved in the apoptotic effect of orexins mediated by the OX1 receptor.
The tyrosine phosphatase SHP-2 mediates orexin-induced apoptosis in the recombinant CHO OX1R-expressing cells
To further document the role of ITIM in the orexin-induced apoptosis mediated by the OX1R, we evaluated whether this GPCR functions as immune receptors do, by recruiting a phophatase of the SH2 domain-containing phosphatase type (28)
. For this purpose, we transfected well-characterized dominant-negative mutants of the two cytoplasmic tyrosine phosphatases SHP-1 and SHP-2 (23)
in CHO/hOX1R WT cells. As shown in Fig. 5
A, B, orexin-A promoted similar extent of apoptosis (Fig. 5A
) and cell growth inhibition (Fig. 5B
) in control cells and in cells transfected with dn-SHP-1, whereas it was not able to induce significant apoptosis (Fig. 5A
) or to inhibit cell growth (Fig. 5B
) in cells transfected with dn-SHP-2. The control transfection of wild-type cDNA of SHP-2 in recombinant CHO OX1R-expressing cells had no effect on orexin-mediated apoptosis or on orexin-mediated cell growth inhibition (data not shown). The physical interaction of SHP-2 with the OX1R was demonstrated by immunoprecipitation experiments (Fig. 5C
). When CHO/hOX1R WT cells were challenged for 5 min with orexin-A, immunoblotting with anti-SHP-2 antibodies of cell lysates immunoprecipitated with anti-OX1R antibodies revealed a coimmunoprecipitation of OX1R and SHP-2. In contrast, little coimmunoprecipitation, if any, could be detected in the absence of orexin-A challenge, supporting the idea that agonist interaction with OX1R is responsible for SHP-2 recruitment. When tyrosine was mutated in the ITIM of the OX1R, challenge of cells with orexin-A did not result in the coimmunoprecipitation of SHP-2 and OX1R, supporting that the physical interaction occurs at the ITIM level. Similar experiments were done with anti-SHP-1 antibodies clearly showing that SHP-1 does not physically interact with the OX1R (Fig. 5B
) in good agreement with the functional data (Fig. 5A
).
Orexin-induced apoptosis involves Y358 phosphorylation of ITIM in OX1R
To further document the role of ITIM in the orexin-induced apoptosis mediated by the OX1R, we evaluated the ability of orexin-A to promote tyrosine phosphorylation in OX1R. When CHO/hOX1R WT cells were challenged for 5 min with orexin-A, immunoblotting with antiphosphotyrosine antibodies of cell lysates immunoprecipitated with anti-OX1R antibodies revealed a dramatic increase of tyrosine phosphorylation. When Y358 was mutated into phenylalanine, no tyrosine phosphorylation could be detected on challenge of cells with orexin-A, supporting that orexin-A induced tyrosine phosphorylation on Y358 in the ITIM of the OX1R. When CHO/hOX1R WT cells were pretreated for 24 h with the Src-like tyrosine kinase inhibitor PP-2, orexin-A no longer promoted tyrosine phosphorylation in OX1R (Fig. 6
A). In line with this latter observation, we demonstrated that the PP2 inhibitor totally abolished orexin-A-induced apoptosis (Fig. 6B
). Altogether, these data suggested that a Src-like tyrosine kinase is involved in orexin-induced tyrosine phosphorylation of Y358 in the OX1R ITIM.
The tyrosine phosphatase SHP-2 mediates orexin-induced apoptosis in the colon cancer cell line HT-29
Since all of the experiments described above were carried out in cell models expressing recombinant OX1R, we extended our work in order to demonstrate that protein tyrosine phosphatases are also instrumental in human colon cancer cells, which express native receptors and in which the proapoptotic action of orexins makes sense in the context of reduction of tumor growth (6)
. The protein tyrosine phosphatase inhibitor PAO (1 µM) totally reversed the orexin-A-induced apoptosis in the human colon adenocarcinoma cells HT-29 (Fig. 7
A). Quantitation indicated that 18.8 ± 0.8% of cells were apoptotic on orexin-A (1 µM) treatment vs. 3.2 ± 0.6% in control cells. In the presence of 1 µM PAO, the proapoptotic action of orexin-A (1 µM) was completely reversed (Fig. 7A
) with a proportion of apoptotic cells of 3.6 ± 0.3%. This phosphotyrosine phosphatase is probably SHP-2, since transfection of HT-29 cells with dnSHP-2 cDNA resulted in a complete inhibition of orexin-A-induced apoptosis (Fig. 7B
).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 7)
. Here, we deciphered the proximal mechanism that couples activation of the membrane-bound G protein-coupled OX1 receptor to cell apoptosis, highlighting an original mechanism (Fig. 8
), which passed unnoticed so far. We demonstrate that an ITIM, a common signature of immune inhibitory receptors (22)
, is present in the OX1R and plays a pivotal role in the OX1R-driven apoptosis by recruiting the protein tyrosine phosphatase SHP-2; and that the OX1R-driven apoptosis requires the presence of the protein Gq but is independent of the activation of phospholipase C and induction of calcium transients by orexins.
|
The OX1R is equipped with a consensus ITIM sequence in the intracellular domain connecting the seventh transmembrane helix to the C-terminal tail (Fig. 8)
. Two sets of observations support a crucial role of this ITIM in the orexin-induced apoptosis. Mutation of the central tyrosine of the ITIM in OX1R completely abolished the ability of orexins to promote apoptosis (Fig. 1)
. This is in agreement with the general paradigm of ITIM function in immune receptors (22)
, in which ligand engagement results in ITIM phosphorylation by Src-like kinases. In this context, we demonstrated that PP2, an inhibitor of Src-like kinases, also abolished orexin-induced apoptosis in CHO cells transfected with the WT OX1R by inhibiting the orexin-induced Y358 phosphorylation in the ITIM (Fig. 6)
, and the orexin-induced apoptosis is blocked by the tyrosine phosphatase inhibitor PAO (Fig. 4)
and also by transfection of a dominant-negative mutant of the phosphotyrosine phosphatase SHP-2 in CHO cells expressing recombinant OX1R and in human colon cancer HT29 cells. The recruitment of SHP-2 by the OX1R was directly demonstrated by coimmunoprecipitation of OX1R and SHP-2 on orexin challenge (Fig. 5C
). Again, these data agree with the paradigm of ITIM function in immune cells since the phosphorylated ITIM recruits phosphotyrosinephosphatases SHP-1 or SHP-2 or the inositol-phosphatase SHIP, which are thereby activated (23)
. The involvement of inositol phosphatase in orexin-induced apoptosis can be ruled out here since PAO is specific of tyrosine phosphatases (29)
and CHO cells do not express any SHIP isoform (30)
, the expression of which is restricted to myeloid lineages (31)
. The recruitment and activation of SHP-2 by OX1R on orexin binding appear to be specific, since the dominant-negative mutant of SHP-1 does not alter orexin-induced apoptosis (Fig. 5)
in transfected CHO cells, which express both forms of SHP (32)
. The same orexin-mediated recruitment of SHP-2 was also demonstrated in HT-29 colon cancer cells that expressed the native OX1R. The role of SHP-2 in promoting apoptosis has been previously noted in embryonic fibroblasts (33)
, primary T helper cells (34)
, and murine erythroleukemia HCD57 cells (35)
. At this stage, the key targets of SHP-2 leading to cell apoptosis on orexin activation of OX1R remain elusive, as well as the targets of SHPs in general (36)
. We have previously demonstrated that orexins promote apoptosis by inducing cytochrome c release into cytosol and activation of caspase-3 and caspase-7 (6)
. A working hypothesis is that on activation of OX1R by orexins, SHP-2 is activated, resulting in the dephosphorylation of signal transducer and activator of transcription (STAT) -5 and its subsequent inactivation, a process leading to a reduction in the expression of the downstream targets of STAT-5, such as antiapoptotic proteins Bcl-XL or the serine threonine kinase Pim-1. This hypothesis is currently under investigation in our laboratory, inasmuch as inactivation of STAT5 by SHP-2 (37)
and the role of STAT5 in controlling the expression of antiapoptotic proteins (37)
have been described in hematopoietic cells. Other hypothesis do exist, since apoptosis in immune cells expresssing immunoreceptors has been shown to be dependent on decreased recruitment/activity of the Ser/Thr kinase Akt (38
39
40)
and/or decreased activity of MAP kinases (41
42
43)
. It is clear that recruitment and activation of SHP-2 by OX1R is such a proximal point of apoptosis regulation that many subsequent apoptosis-related events have to be investigated in future studies.
Since OX1R is known to be coupled to the Gq protein and on activation it elevates cytosolic calcium level (reviewed in ref. 3
), an important issue was to determine the role of ITIM on the one hand and of Gq and orexin-induced calcium transients on the other hand in the apoptotic response. We observed that mutation of the central tyrosine of ITIM into phenylalanine, uncouples OX1R from Gq. This was based on several observations, including the shift of mutated receptor affinity for orexin (Fig. 1)
, the inability of Gpp(NH)p to inhibit labeled orexin binding to the mutated receptor (Fig. 2)
, and the inability of orexins to induce calcium transients in CHO cells expressing the mutated receptor (Fig. 2)
. Although the integrity of ITIM appears to be necessary for coupling of OX1R to Gq, it is clear that some serpentine receptors, which are not equipped with ITIM, such as the NTS1 receptor for neurotensin (12)
or the protease-activated receptor 1 for thrombin (13)
, do exhibit efficient coupling to Gq. This difference can be tentatively ascribed to the fact that the location of the ITIM is a site of interaction of Gq in some receptors (27)
but not in others. To understand the exact role of Gq in the orexin-induced apoptosis, we transfected the WT OX1R in mouse embryonic fibroblast cells, in which the GQ subunit was invalidated (26)
. The presence of Gq was clearly shown to be necessary for the ability of orexins to induce apoptosis (Fig. 3)
. In the meantime, we demonstrated that even though activation of OX1R by orexins recruits Gq resulting in the stimulation of phospholipase C as evidenced by stimulation of inositol phosphate formation (Fig. 2)
and induction of calcium transients (Fig. 2)
, the activation of phospholipase C is not involved in the orexin-induced apoptotic response. Indeed, the phospholipase C inhibitor U-73122 completely reversed the effect of orexins on inositol phosphate formation and calcium transients but did not alter the apoptotic response. It follows that orexin-induced calcium transients (Fig. 2)
is not necessary for the orexin-induced apoptotic response, whatever the origin of calcium either from intracellular stores through an InsP3-driven mechanism (25)
or by activation of a Ca2+ entry pathway that involves diacylglycerol-activated TRPC channels (8
, 44)
. Of course, that does not preclude basal cytosolic calcium from playing a role in the onset of apoptosis (13)
. Because the classical role of Gq in activating phospholipase C is not involved in orexin-induced apoptosis, the question is to know what is the role of Gq in this process. Although not addressed experimentally in this work, it is likely that on orexin binding to OX1R, the heterotrimeric Gq protein dissociates, releasing not only the q subunit but also the β dimer (45)
. The availability of β dimers could represent a crucial link between Gq and the orexin-induced apoptosis driven by the ITIM in the OX1R. Indeed, the tyrosine kinase c-Src, which phosphorylates the central tyrosine of ITIM (19)
, is activated by free β dimers (46)
. In this context, it can be hypothesized that on orexin binding to OX1R, the Gq protein, is activated releasing β dimers that stimulate Src-like enzymes allowing phosphorylation of the tyrosine of ITIM. On tyrosine phosphorylation, OX1R can recruit the SHP-2 phosphotyrosine phosphatase and activates thereby SHP-2, initiating apoptosis as discussed above. This mechanism is schematized in Fig. 8
and accounts for all the data shown in this paper, including the dual participation of Gq and ITIM in the orexin-induced apoptotic response.
In conclusion, the present work sheds new light on the mechanism whereby a G protein-coupled receptor drives agonist-induced apoptosis. This work has important implications in two domains. First, it provides a mechanism accounting for the dramatic action of orexins in the induction of apoptosis. Since OX1R is expressed in most human colon cancer cell lines (6)
and resected tumors (unpublished results) but not in normal human colonic epithelial cells (6)
, the aberrant expression of OX1R in colon cancer represents an attractive target for the development of instrumental OX1R agonists in this cancer. Such OX1R agonists would have the great advantage to promote apoptosis in colonic tumor cells and not in normal colonic epithelial cells (6)
. In that respect, OX1R represents a more attractive target in colon cancer therapy than downstream enzymes, including SHP-2, which are expressed in both normal and cancerous epithelial cells (47
, 48)
. A second important implication of the present work is related to the fact that very few reports describe a role for ITIM in G protein-coupled receptors, in all cases, in relation to the control of cell proliferation (32)
. The present work adds a new dimension to the role of ITIMs in GPCRs, e.g., induction of apoptosis. Moreover, mining the GPCR sequence databases for the presence of ITIMs (unpublished results) revealed that ITIMs are much more frequent (25–30%) in GPCRs than initially thought, with widespread expression in some subclasses of GPCRs, such as odorant receptors, 80% of which are equipped with an ITIM. This emphasizes the need to reinterpret the signal transduction of many GPCRs by taking into account the role of ITIMs in apoptosis (present work), proliferation (32)
, and possibly other effects.
|
ACKNOWLEDGMENTS |
|---|
Received for publication September 14, 2007. Accepted for publication December 13, 2007.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
family of heterotrimeric G proteins. Cell Signal. 18,135-150[CrossRef][Medline] proteins and c-SRC tyrosine kinase in parathyroid hormone-induced signal transduction in rat enterocytes. J. Endocrinol. 188,69-78This article has been cited by other articles:
|
|
 |
|
  A. El Firar, T. Voisin, C. Rouyer-Fessard, M. A. Ostuni, A. Couvineau, and M. Laburthe Discovery of a functional immunoreceptor tyrosine-based switch motif in a 7-transmembrane-spanning receptor: role in the orexin receptor OX1R-driven apoptosis FASEB J, December 1, 2009; 23(12): 4069 - 4080. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Orabona, M. T. Pallotta, C. Volpi, F. Fallarino, C. Vacca, R. Bianchi, M. L. Belladonna, M. C. Fioretti, U. Grohmann, and P. Puccetti SOCS3 drives proteasomal degradation of indoleamine 2,3-dioxygenase (IDO) and antagonizes IDO-dependent tolerogenesis PNAS, December 30, 2008; 105(52): 20828 - 20833. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
) cells on a 48-h treatment with the peptide in the presence of 5% FCS. B) Dose-response of orexin-A in inhibiting cell growth in CHO/hOX1R WT (•) and CHO/hOX1R Y358F (
