Molecular Endocrinology, doi:10.1210/me.2004-0389
Molecular Endocrinology 20 (1): 80-99
Copyright © 2006 by The Endocrine Society
OX1 Orexin Receptors Activate Extracellular Signal-Regulated Kinase in Chinese Hamster Ovary Cells via Multiple Mechanisms: The Role of Ca2+ Influx in OX1 Receptor Signaling
Sylwia Ammoun,
Lisa Johansson,
Marie E. Ekholm,
Tomas Holmqvist,
Alexander S. Danis,
Laura Korhonen,
Olga A. Sergeeva,
Helmut L. Haas,
Karl E. O. Åkerman and
Jyrki P. Kukkonen
The Department of Neuroscience (S.A., L.J., M.E.E., T.H., A.S.D., K.E.O.Å., J.P.K.), Unit of Physiology, Uppsala University, Biomedical Center (BMC), SE-75123 Uppsala, Sweden; the Department of Neuroscience (L.K.), Unit of Neurobiology, Uppsala University, BMC, SE-75123 Uppsala, Sweden, and Minerva Foundation Institute for Medical Research, Helsinki, Finland; Department of Neurophysiology (O.A.S., H.L.H.), Heinrich-Heine-Universität, D-40001 Düsseldorf, Germany; and the A. I. Virtanen Institute for Molecular Sciences (K.E.O.Å.), University of Kuopio, Neulaniementie 2, FIN-70210 Kuopio, Finland
Address all correspondence and requests for reprints to: Jyrki P. Kukkonen, Department of Neuroscience, Division of Physiology, Uppsala University, BMC, P.O. Box 572, SE-75123 Uppsala, Sweden. E-mail: jyrki.kukkonen{at}neuro.uu.se.
 |
ABSTRACT
|
|---|
Activation of OX1 orexin receptors heterologously expressed in Chinese hamster ovary (CHO) cells led to a rapid, strong, and long-lasting increase in ERK phosphorylation (activation). Dissection of the signal pathways to ERK using multiple inhibitors and dominant-negative constructs indicated involvement of Ras, protein kinase C, phosphoinositide-3-kinase, and Src. Most interestingly, Ca2+ influx appeared central for the ERK response in CHO cells, and the same was indicated in recombinant neuro-2a cells and cultured rat striatal neurons. Detailed investigations in CHO cells showed that inhibition of the receptor- and store-operated Ca2+ influx pathways could fully attenuate the response, whereas inhibition of the store-operated Ca2+ influx pathway alone or the Ca2+ release was ineffective. If the receptor-operated pathway was blocked, an exogenously activated store-operated pathway could take its place and restore the coupling of OX1 receptors to ERK. Further experiments suggested that Ca2+ influx, as such, may not be required for ERK phosphorylation, but that Ca2+, elevated via influx, acts as a switch enabling OX1 receptors to couple to cascades leading to ERK phosphorylation, cAMP elevation, and phospholipase C activation. In conclusion, the data suggest that the primary coupling of orexin receptors to Ca2+ influx allows them to couple to other signal pathways; in the absence of coupling to Ca2+ influx, orexin receptors can act as signal integrators by taking advantage of other Ca2+ influx pathways.
|
INTRODUCTION
|
|---|
OREXINS, ALSO CALLED hypocretins, are recently discovered peptide signal substances (1, 2), that act via two G protein-coupled receptors (GPCRs), OX1 and OX2 receptors. Orexins (orexin-A and -B) may act both as synaptic transmitters in the central and peripheral nervous systems and as para- and endocrine mediators in the central nervous system (CNS) and peripheral organs (reviewed in Ref.3). Action of orexins as para-/endocrine mediators is suggested by circumstantial evidence such as putative nonsynaptic release sites in the CNS and the lack of orexinergic innervation of many orexin receptor-expressing and orexin-responding peripheral tissues such as muscle cells of the gastrointestinal tract, endocrine cells of the gastrointestinal tract, pituitary gland, adrenal gland (cortex and medulla), and testis (Ref.4 ; reviewed in Ref.3), and hematopoietic cells (5, 6). Furthermore, pancreatic 
- and ß-cells, endocrine (enterochromaffin) cells of the gastrointestinal tract, adrenal and pituitary glands, and testis may release orexins, and blood plasma may contain orexin from some peripheral source (Refs.7 and 8 ; reviewed in Ref.3). Orexins regulate 1) vigilance and sleep pattern, 2) feeding, appetite, and metabolic processes and 3) hormonal responses, both via CNS and other organ systems, in particular the gastrointestinal tract and endocrine glands (reviewed in Ref.3).
Signal coupling of orexin receptors has been rigorously investigated in few studies. One of the most marked responses to orexin receptor activation is a robust Ca2+ influx in native neurons (reviewed in Ref.3), in the STC-1 endocrine cell line (9) and, upon heterologous receptor expression, in neuron/neuroendocrine-like cells (10) and Chinese hamster ovary (CHO) cells (11, 12). In addition to Ca2+ influx, phosphatidylinositol-specific phospholipase C (PLC)-dependent Ca2+ release is activated in orexin receptor-expressing CHO cells (11); PLC activation is also seen in endocrine cells of adrenal gland (medulla and cortex?) and testis (4, 13, 14). Production of cAMP in response to orexin receptor activation has been observed in adrenal cortical and CHO cells and in hypothalamic neurons (13, 14, 15, 16). The mechanisms involved in the orexin-triggered signals are largely unknown. OX2 receptors have been suggested to be able to couple to Gs, Gi/o and Gq proteins (14), but the roles of different G proteins in physiological orexin receptor signaling are unknown.
Thus, some responses are common for different cell types, whereas some have only been demonstrated in specific cell types. Differences may derive from divergent expression of signal molecules (e.g. adenylyl cyclase isoforms) in different cell types, but most observed differences should yet be resolved by more focused studies: for instance, native neurons have not consequently been investigated for PLC or adenylyl cyclase activation whereas endocrine cells have not been subject to detailed Ca2+ measurements. Cell type-specific variation in orexin receptor signaling thus remains to be shown. It appears, though, that there are general signaling mechanisms for orexin receptors, because some of the signal pathways, e.g. activation of Ca2+ influx pathways and PLC, are present independently of expression system.
Many different types of GPCRs have been shown to be able to affect cell fate largely via the same or similar signal cascades as receptor tyrosine kinases (reviewed in Refs.17 and 18). This may be particularly relevant for orexins because orexins are likely to produce long-lasting receptor stimulation through their para-/endocrine action. In particular, orexins have effects on adrenal cortex similar to ACTH (13, 14, 15). We have previously demonstrated that OX1 receptors heterologously expressed in CHO cells are capable of activating ERK phosphorylation (3). This has also been verified in a recent study (19). However, the signal pathways involved in the ERK activation have not been identified in any of these studies. We have therefore in this study investigated the mechanistic basis of the OX1 receptor-mediated ERK activation and its relationship to other known orexin receptor-activated signals. The results demonstrate involvement of multiple signal cascades in this response and suggest a novel signaling scheme, in which Ca2+ influx is required as a switch allowing OX1 receptor to couple to other signal cascades.
|
RESULTS
|
|---|
Orexin Receptor Activation Leads to ERK Phosphorylation
In CHO cells recombinantly expressing OX1 or OX2 receptors, orexin receptor stimulation strongly increased levels of phosphorylated ERK (pERK)1 and -2 (Fig. 1
, A and B). The exact maximum magnitude is hard to assess because the basal (nonstimulated) level was often below the detection limit, but in those cases in which it was possible to measure, the stimulated level of pERK2 was 6.1 ± 0.7 times the basal (six independent experiments). Phosphorylation of ERK1 and -2 followed the same pattern. More detailed investigations were conducted with cells expressing OX1 receptors. The elevation in pERK reached its maximum in 3 min and declined very slowly (Fig. 1C). Yet continuous OX1 receptor activity was required as ERK phosphorylation declined when the receptors were blocked by the OX1 receptor-selective antagonist, SB-334867 (1-[2-methylbenzoxazol-6-yl]-3-[1,5]naphthyridin-4-yl-urea hydrochloride; Fig. 1D). The responses to orexin-A and -B also displayed a clear concentration dependency (Fig. 2A) with the expected potency order (orexin-A 18-fold more potent than orexin-B). We compared the potencies of the orexin peptides with respect to ERK phosphorylation to a more well-known orexin response in CHO cells, Ca2+ elevation (Fig. 2B; the Ca2+ data are derived from Ref.20). OX1 receptor activation elevates Ca2+ via two mechanisms, a pure release (seen in the absence of extracellular Ca2+; solid symbols) and a receptor-operated Ca2+ influx-dependent elevation (seen in the presence of extracellular Ca2+; open symbols) (11, 12, 21). The absolute potencies of the orexins toward the ERK response were approximately 10-fold lower as compared with their potencies toward Ca2+ elevation (compare Fig. 2A, open symbols, to Fig. 2B, open symbols). As expected, OX1 receptor-stimulated ERK phosphorylation was fully blocked by the inhibition of 1) MKK1 (MEK1, MAPK/ERK kinase 1) activation using U0126 (1,4-diamino-2,3-dicyano-1,4-bis[o-aminophenylmercapto]butadiene) and 2) MKK1 activity using dominant-negative MKK1 (data not shown).
View larger version (24K):
[in this window]
[in a new window]
|
Fig. 1. Orexin Receptors Stimulate ERK Phosphorylation in Recombinant CHO Cells
A, OX1 receptors; B, OX2 receptors. C, OX1 receptor-expressing cells were exposed to stimulation with orexin-A for the given times. D, OX1 receptor-expressing cells were stimulated with 100 nM orexin-A for 30 min. Solid circles indicate cells pretreated with the OX1 receptor antagonist SB-334867 (10 µM) for 30 min before exposure to orexin-A (SB-334867 was also present throughout the experiment). Open triangles indicate cells to which SB-334867 was added after 10 min of orexin-A stimulation. Control cells (open circles) were exposed to orexin-A only. ctrl, Control.
|
|
View larger version (19K):
[in this window]
[in a new window]
|
Fig. 2. Concentration-Response Curves for ERK Phosphorylation (A) and Ca2+ Elevation (B) in OX1 Receptor-Expressing CHO Cells
A, The solid lines indicate the best fits for the data: for orexin-A, EC50 = 21 ± 4 nM, nHill (Hill-coefficient) = 1.4 ± 0.3 (P < 0.05 as compared with 1); for orexin-B, EC50 = 380 ± 100 nM, nHill = 1.The apparent cooperativity (nHill) for orexin-B was not significantly different from 1. B, Ca2+ elevation ( [Ca2+]i =change in intracellular [Ca2+] = [Ca2+]i/stimulated – [Ca2+]i/basal) by orexin peptides in 1 mM (+Ca2+; open symbols) and 140 nM (–Ca2+; filled symbols) extracellular Ca2+ as measured in suspensions of CHO cells. Derived from Ref.20 . The gray vertical lines in each subfigure mark peptide concentration of 100 nM for comparison.
|
|
ERK Phosphorylation by OX1 Receptors Is Highly Dependent on Ca2+ Influx
Ca2+ elevation, especially influx, appears to be central for orexin signaling in neurons and in recombinant cells (see Introduction and Discussion). OX1 receptor signaling to Ca2+ elevation in CHO cells is highly dependent on extracellular Ca2+ (Ref.11 ; see also above). Extracellular Ca2+ does not affect orexin binding (2, 11), but it is required for influx of Ca2+ into cells via a receptor-operated (and other?) Ca2+ influx pathway(s) (11, 21). This and the fact that Ca2+ influx through native Ca2+ channels has been shown to be important for the activation of ERK, as well as other responses, in different types of cells (22, 23, 24, 25), prompted us to investigate the role of Ca2+ in ERK activation via OX1 receptors. When we reduced the extracellular Ca2+ to 140 nM, a concentration that is low enough to prevent Ca2+ influx (11, 21), the orexin receptor-mediated phosphorylation of ERK was fully attenuated (Fig. 3, A and B). A similar behavior was seen in heterologous OX1 receptor-expressing neuro-2a cells (Fig. 3C). Primary cultured striatal neurons (Fig. 3D) displayed low basal pERK levels, which were markedly elevated by orexin-A in a subset of cells. Addition of EGTA before orexin-A eliminated the increase in pERK staining (Fig. 3D). Investigations of primary neuronal cultures are much complicated by the heterogeneity of the cells, i.e. lack of orexin receptor expression in large cell populations, secondary responses due to transmitter release, and difficult transfections. We therefore decided to proceed with more detailed investigations of the mechanisms in CHO cells.
View larger version (28K):
[in this window]
[in a new window]
|
Fig. 3. Dependence of Orexin Receptor-Mediated ERK Phosphorylation on Extracellular Ca2+
A–D, the effect of reduction of extracellular Ca2+ from 1 mM (+Ca2+) to 140 nM (–Ca2+) with 0.9 mM EGTA on the orexin and thapsigargin response in CHO-hOX1 (A and B) and neuro-2a-hOX1 (C) cells and in primary cultures of striatal neurons (D). The cells were stimulated with orexin-A or thapsigargin for 10 (CHO cells and primary neurons) or 30 min (neuro-2a cells). EGTA was added 1 min before the stimulation. The data in B and C are normalized to the basal pERK level (0%) and the maximum orexin-A stimulation (100%) in 1 mM Ca2+. D, The concentration of orexin-A is 100 nM. E, CHO-hOX1 cells were stimulated with 1 µM thapsigargin or 1 µM ionomycin for the indicated times in 1 mM extracellular Ca2+. F, CHO-hOX1 cells were stimulated with 100 nM orexin-A for the indicated times. Solid circles designate cells pretreated with EGTA for 1 min before the exposure to orexin-A (EGTA was also present throughout the experiment). Open triangles designate cells to which EGTA was added after 3 min of orexin-A stimulation. Ctrl cells (open circles) were exposed to orexin-A only.
|
|
We next investigated whether Ca2+ elevation itself could activate ERK phosphorylation. Intracellular Ca2+ levels were elevated using thapsigargin, an endoplasmic/sarcoplasmic reticulum Ca2+ ATPase (SERCA) inhibitor, which releases Ca2+ from the endoplasmic reticulum, hence activating Ca2+ influx via the so-called store-operated Ca2+ channels, and the Ca2+ ionophore ionomycin, which permeabilizes the cells for Ca2+. Both thapsigargin and ionomycin increased ERK phosphorylation, but to a lower extent than orexins, and their action declined much faster (Fig. 3E
), despite persisting Ca2+ influx (not shown for ionomycin; for thapsigargin, see Fig. 4C). Even the ERK response (only tested for thapsigargin) was fully reversed by reduction in extracellular [Ca2+] (Fig. 3B).
View larger version (20K):
[in this window]
[in a new window]
|
Fig. 4. OX1 Receptor Stimulation and Thapsigargin Cause Long-Lasting Ca2+ Response
A, Ca2+ responses in 10 individual CHO cells stimulated with 100 nM orexin-A for 10 min. B and C, Averaged responses of individual CHO cells stimulated with orexin-A (B) or thapsigargin (C) over an 80-min time period. Ca2+ removal-Ca2+ readdition protocol was applied to better visualize the continuous Ca2+ influx. The traces represent averages of about 20 cells. Standard error is shown for every tenth point. s, Second.
|
|
Reduction of extracellular [Ca2+] in CHO cells during orexin receptor stimulation also restored the basal ERK phosphorylation state (Fig. 3F; compare with Fig. 1D with addition of SB-334867). Thus, reduction of extracellular [Ca2+] reversed orexin-A-induced ERK phosphorylation independently of whether Ca2+ was lowered before or after orexin-A. Together with the fact that Ca2+ elevation via OX1 receptors at 
100 nM orexin-A was unaffected by reduction of extracellular Ca2+ (Fig. 2B; compare empty and solid circles), this suggests that Ca2+ influx, rather than Ca2+ elevation as such, is essential for the ERK activation. The experiment in Fig. 3F also suggests that Ca2+ influx not only primes the ERK phosphorylation but is also required to maintain it. Therefore, OX1 receptor activation should lead to long-lasting Ca2+ influx. When CHO cells were exposed to 100 nM orexin-A continuously, the Ca2+ signal was apparently strongly attenuated in most cells after only a few minutes (Fig. 4A). Two possible explanations for this are desensitization of orexin receptor signaling or high cellular Ca2+ pump activity that hides the continuous Ca2+ influx from fura-2. We scrutinize these alternatives in Fig. 4B. After 1-h exposure to orexin-A the intracellular [Ca2+] was apparently only weakly elevated (Fig. 4B). This elevation was nevertheless significant, and it was more clearly visualized upon lowering of extracellular [Ca2+]. When extracellular [Ca2+] of 1 mM was restored still in the presence of orexin-A, a much higher intracellular Ca2+ level was reached. Thus, reduction of extracellular [Ca2+] prevents Ca2+ influx allowing Ca2+ pumps to attain a less active state. Upon restoration of normal extracellular [Ca2+] concentration, an apparently higher Ca2+ level is obtained because the Ca2+ pumps are not activated as rapidly as the influx. Thus, these data show that OX1 receptor activity, and Ca2+ influx, is substantial after a 1-h exposure to orexin-A, although Ca2+ pump activity is hiding a significant part of the Ca2+ flux. The absolute level of Ca2+ in the cellular compartment likely to be most relevant for the signaling in this case, the submembrane compartment (see below), cannot be estimated because fura-2 measures overall cellular Ca2+ levels; however, it can be safely assumed to be significantly higher than that suggested by the fura-2 recording. Thapsigargin-induced Ca2+ influx behaved in the same way as the orexin-induced influx (Fig. 4C).
In addition to activating receptor-operated Ca2+ channels, OX1 receptors also activate store-operated Ca2+ influx via inositol-1,4,5-trisphosphate (IP3) -dependent Ca2+ release (21). To determine which channels were involved in which particular response, we used blockers of these channel types. Orexin-A (1–3 nM) was used for Ca2+ measurements, because the Ca2+ responses at these concentrations are fully dependent on the receptor-operated influx (e.g. see Fig. 2B) and can thus be used as a measure for the receptor-operated influx. Orexin-A (100 nM) was used for ERK measurements. Thapsigargin, which only activates the store-operated pathway, was used as a control for this influx pathway (21). 2-APB (2-aminoethoxydiphenyl borate), which blocks most of the store-operated Ca2+ channel activity in CHO cells (Refs.21 and 26 ; reviewed in Ref.27), apparently fully abolished the thapsigargin-induced Ca2+ influx and ERK phosphorylation (Fig. 5, B–D) but had no significant effect on the peak Ca2+ (dependent on receptor-operated influx) or ERK responses to orexin-A (Fig. 5, A and C–D). The nonselective cation channel blocker Ni2+ also blocked the Ca2+ influx response to thapsigargin (Fig. 5, B and C) and, to a large extent, the orexin receptor-operated Ca2+ influx seen at 1–3 nM orexin-A (Fig. 5, A and C; see also Ref.21). To an equal degree as it inhibited Ca2+ influx responses to thapsigargin and orexin-A, Ni2+ also inhibited the ERK phosphorylation (Fig. 5D). These data suggest that, at least in part, separate calcium channels, orexin receptor-operated and store-operated, are involved in the ERK activation by OX1 receptors and thapsigargin, respectively.
View larger version (30K):
[in this window]
[in a new window]
|
Fig. 5. Ca2+ Channel Types Involved in the ERK Response to Orexin-A and Thapsigargin (D) Compared with the Overall Ca2+ Channel Activity during the Same Treatments (A–C) in OX1 Receptor-Expressing CHO Cells
2-APB was used to block store-operated Ca2+ influx pathways, and Ni2+ to block both receptor- and store-operated pathways (21 26 ). Peak response to low concentration of orexin-A (0.3–3 nM; A and C) can be used as a selective measure for the receptor-operated influx, whereas the peak response at higher concentrations (100 nM) is determined by influx-independent Ca2+ release (Refs.11 12 21 and 26 ; see also Fig. 2B). The inhibitors can be applied either during the stimulus (as shown for thapsigargin in B) or before (5 min) stimulus (as shown for orexin-A in A) with similar results. A and B, Traces of averaged single-cell measurements; C, statistics of the responses as measured in cell suspension. Standard error is shown for every third (A) or tenth (B) point. In C, open bars represent orexin-A (3 nM) peak response, dependent on the receptor-operated Ca2+ influx, and solid bars represent the thapsigargin-mediated (1 µM) Ca2+ influx at the stabilized level (2 min after addition), solely relying on the store-operated Ca2+ fluxes (21 26 ). The comparison for each inhibitor is to the control orexin-A or thapsigargin response. Control responses have error bars because the responses were averaged from the absolute Ca2+ levels. D, ERK phosphorylation was measured after 10-min exposure to 100 nM orexin-A or 3-min exposure to 1 µM thapsigargin. 2-APB and NiCl2 were added 5 min before orexin or thapsigargin. The comparison for each inhibitor is to the control orexin-A or thapsigargin response. Control responses have no error bars because the inhibited responses have been normalized to the controls in each different batch of cells. ctrl, Control; s, second; ns, not significant.
|
|
Diacylglycerol-Dependent Protein Kinase C (PKC) Activity, But Not Ca2+ Release, Contributes to the ERK Phosphorylation by OX1 Receptors
As indicated above, OX1 receptors also connect to PLC and therewith to generation of IP3, as we have previously shown with biochemical measurements of IP3 and total inositol phosphates (11), and putatively IP3-dependent Ca2+ release (Ref.11 ; see also Fig. 2B of the present study). We further confirmed this by measuring translocation of green fluorescent protein (GFP)-linked PH-domain of PLC
1 (GFP-PH-PLC1) and GFP-linked PKC
(GFP-PKC) upon orexin receptor activation. The former rapidly translocated from the plasma membrane to the cytosol (Fig. 6A), suggesting breakdown of PIP2 (phosphatidylinositol-4,5-bisphosphate)/generation of IP3, whereas the latter translocated from the cytosol to the plasma membrane (Fig. 6B), suggesting generation of diacylglycerol. Thus, orexin receptor stimulation strongly activates PLC. To investigate the role of the PLC signaling in the ERK phosphorylation, we exposed the cells to 3 µM of the PLC inhibitor U-73122 (1-[6-([(17b)-3-methoxyestra-1,3,5(10)-trien-17-yl]amino)hexyl]-1H-pyrrole-2,5-dione; 10 min preincubation). This concentration fully blocked the inositol phosphate generation and Ca2+ release (data not shown). ERK phosphorylation in response to orexin-A stimulation was reduced by about 60% (Fig. 7A). Because PLC generates both inositol phosphates (e.g. IP3) and diacylglycerol (activator of e.g. conventional and novel isoforms of PKC), we wanted to investigate which of these messengers was important for OX1 receptor signaling to ERK. Direct stimulation of conventional and novel isoforms of PKC using TPA (12-O-tetradecanoylphorbol 13-acetate) caused a marked increase in ERK phosphorylation (Fig. 7, B and C). PKC can be inhibited using pharmacological inhibitors or by down-regulating its activity via prolonged treatment of the cells with a PKC activator. We used both of these approaches to judge whether PKC was playing a part in the orexin receptor signal to ERK. Preincubation for 30 min with 1 µM of the specific PKC inhibitor GF109203X (=bisindolylmaleimide I = Gö6850 = 2-[1-(3-dimethylaminopropyl)-1H-indol-3-yl]-3-(1H-indol-3-yl)-maleimide) or treatment for 24 h with 1 µM of the PKC activator TPA fully blocked the ERK response to subsequent addition of TPA (data not shown). Responses to orexin-A were also significantly, but not completely (by 
50%), reduced by either treatment (Fig. 7D). The concentration of GF109203X chosen, 1 µM, should relatively selectively affect conventional and novel isoforms of PKC (16, 28). A selective inhibitor of conventional PKCs, Gö6976 (12-[2-cyanoethyl]-6,7,12,13-tetrahydro-13-methyl-5-oxo-5H-indolo[2,3-a]pyrrolo[3,4-c]-carbazole; 1 µM) (16, 28), produced an inhibition of an equal degree as that by GF109203X and long-term TPA treatment (Fig. 7D), suggesting that conventional PKC(s) mediate this response. This is in contrast to stimulation of adenylyl cyclase by OX1 receptors, which engages novel PKC in CHO cells (16). As conventional PKCs are Ca2+ sensitive, we investigated whether there would be synergistic effects between Ca2+ elevation and PKC activation using thapsigargin/ionomycin and submaximal doses of TPA. The results fail to show any synergism (Fig. 7E). Most interestingly, the inhibition of the orexin response produced by GF109203X, Gö6976, and long-term TPA treatment is of the same magnitude as the inhibition by U-73122 (compare Fig. 7, panels D and A). Measurements in the same batches of cells underlined this further: the response to orexin-A in the presence of GF109203X was equal to the response in the presence of U-73122 (GF109203X 3.9 ± 1.7% higher, not significant). This suggests that of the signals generated via PLC, diacylglycerol mobilization and subsequent PKC activation, but not IP3 or Ca2+ release, are important for OX1 receptor signaling to ERK.
View larger version (32K):
[in this window]
[in a new window]
|
Fig. 7. Dependence of OX1-Receptor-Mediated ERK Activation on PLC and Conventional/Novel PKCs in CHO Cells
A, Inhibition of the orexin-A-induced (100 nM) ERK phosphorylation by the PLC inhibitor U-73122 (3 µM, 10 min preincubation). B and C, ERK phosphorylation in response to OX1 receptor stimulation and direct PKC stimulation with TPA. Orexin-A bar in C has no standard deviation because TPA (100 nM, 30 min) response has been normalized to it in each experiment. The comparison is to the orexin response. D, Statistics of the inhibition of PKCs on orexin-A-induced ERK phosphorylation. The cells were preincubated with 1 µM GF109203X or Gö6976 for 30 min or with 1 µM TPA for 24 h before application of 100 nM orexin-A. The dotted lines in A and D indicate the noninhibited orexin response. The comparisons are to the noninhibited response (100%). E, The lack of synergism between Ca2+ elevation (thapsigargin, ionomycin) and PKC activation (TPA). The cells were treated with 1 µM thapsigargin or ionomycin for 3 min and different concentrations of TPA for 30 min. Thus, for coadditions, the cells were stimulated at time 0 with TPA; thapsigargin or ionomycin was added at 27 min and the experiments were interrupted at 30 min. ctrl, Control; +GF, + GF109203X; +Gö, Gö6976.
|
|
The Role of Other Classical Intracellular Signal Pathways in ERK Phosphorylation
Ras is often a central mediator of ERK activation due to its involvement in several cascades to ERK [e.g. direct interaction with Raf, activation of PI3K (phosphoinositide-3-kinase) 
PDK1 (3'-phosphoinositide-dependent kinase 1) PKC]. Ras activation can integrate multiple signals via GEFs (GDP/GTP exchange factors) sensitive to Ca2+ and diacylglycerol and possibly even G protein ß
-subunits (reviewed in Refs.29 and 30). Ras isoforms show different cellular localization (reviewed in Ref.31); they may be activated by different stimuli (e.g. see Ref.32), and they may have different abilities to activate different Ras effector pathways (e.g. see Ref.33). On the other hand, a dominant-negative variant of H-Ras, H-RasS17N, has previously been reported to be an efficient inhibitor of all traditional Ras-isoforms [H-, K-, and N-Ras (34)]. This construct indeed very effectively, but not completely, inhibited orexin-induced ERK activation (Fig. 8A). Similar dominant-negative variants of K- and N-Ras separately were not more effective than H-Ras alone (data not shown). However, expression of all the dominant-negative Ras constructs together almost completely blocked ERK phosphorylation (Fig. 8A). In contrast, dominant-negative variants of the related GTPases RalA and Rap1a did not inhibit OX1 receptor-mediated ERK phosphorylation (Fig. 8A).
View larger version (28K):
[in this window]
[in a new window]
|
Fig. 8. Dependence of OX1-Receptor-Mediated ERK Activation on Ras-Related Small G Proteins (A), Src (B), PI3K(s) (C), and Atypical PKC(s) (D and E) in CHO Cells
A, The cells were transfected with the dominant-negative H-Ras, RalA, or Rap1a or with dominant-negative H-, K- and N-Ras together. B, The cells were pretreated with the Src family kinase inhibitor SU6656 (2 µM) for 30 min before application of 100 nM orexin-A, or they were transfected with the dominant-negative (dn) Src or the Src-inactivating kinase Csk. C, The cells were pretreated with the non-subtype-selective inhibitors of PI3Ks, wortmannin (wortm; 300 nM) and LY294002 (10 µM), for 30 min before application of 100 nM orexin-A, or they were transfected with an activated variant of the PIP3 hydrolyzing enzyme PTEN or dominant-negative p85. D and E, Inhibition of orexin-A-induced ERK phosphorylation by two different dominant-negative PKC constructs. The dotted lines in A, B, C, and E indicate the noninhibited orexin response. The comparisons are to the noninhibited response (100%). ns, not significant.
|
|
The Src family of protein kinases has been implicated in the GPCR signaling to MAPK cascades, although the molecular mechanism is often unclear (reviewed in Ref.18). The involvement of the Src family in the OX1-receptor-mediated ERK phosphorylation was assessed by the pharmacological inhibitor of the Src family of kinases, SU6656 (2,3-dihydro-N,N-dimethyl-2-oxo-3-[(4,5,6,7-tetrahydro-1H-indol-2-yl)methylene]-1H-indole-5-sulfonamide) (2 µM), and by expression of two constructs, the physiological inhibitor of the activation of Src family, the protein kinase Csk, and a dominant-negative Src. A significant inhibition was obtained with either of these inhibitors, although SU6656, for an unclear reason, was significantly (P < 0.05) less efficient (Fig. 8B).
Class I PI3Ks have been implicated in the signaling of GPCRs upon activation via G protein ß-subunits or via Ras and phosphotyrosines (reviewed in Ref.35), and the elevation in PIP3 (phosphatidylinositol-3,4,5-trisphosphate) might lead to activation of PKC (and therewith ERK) via PDK1. The non-subtype-selective PI3K inhibitors wortmannin (300 nM) and LY294002 (2-[4-morpholinyl]-8-phenyl-1[4H]-benzopyran-4-one; 10 µM) and expression of an activated variant of the physiological inhibitor of the PIP3 (phosphatidylinositol-3,4,5-trisphosphate) signaling, the PIP3 3'-phosphatase PTEN (phosphatase and tensin analog), all reduced the orexin signal by about 50–60% (Fig. 8C). A selective inhibitor of Class Ia PI3Ks, dominant-negative p85, also inhibited the response to the same extent. Inhibition of PKC with GF109293X in the presence of wortmannin or dominant-negative p85 did not produce any further inhibition of ERK phosphorylation (not shown), suggesting that PKC resides in the same pathway to ERK as PI3K.
We also observed a significant inhibition of ERK phosphorylation by two different dominant-negative PKC constructs (Fig. 8, D and E), which suggests that atypical PKC isoforms are also involved in the ERK response. Atypical PKCs are not activated by the same stimuli as the conventional and novel isoforms of PKC, except for the phosphorylation by, for instance, PDK1 (e.g. see Ref.36), and they would thus not lie downstream from PLC but rather from, e.g. PI3K or Ras.
We also performed Ca2+ measurements with GF109203X, dominant-negative Ras, dominant-negative Src, wild-type Csk, wortmannin, active PTEN, and dominant-negative PKC to ensure that they did not have nonspecific effects on cell signaling. None of these affected the Ca2+ response to orexin-A, i.e. the number of cells responding, the magnitude of the response, or the kinetics of the response at different concentrations of orexin-A (data not shown). We therefore consider that their effects on OX1 receptor-mediated ERK phosphorylation are specific.
ERK Phosphorylation by OX1 Receptors Is Independent of cAMP Elevation
OX1 receptors expressed in CHO cells strongly elevate cAMP (16), and cAMP-induced ERK phosphorylation has previously been measured in CHO cells (e.g. see Ref.37). We were hence interested in investigating, whether cAMP elevation could be involved in the ERK response to OX1 receptor stimulation. We first tested the activity of adenylyl cyclase inhibitors 2',3'-dideoxyadenosine, MDL-12,330A (cis-N-[2-phenylcyclopentyl]azacyclotridec-1-en-2-amine) and SQ 22536 (9-[tetrahydro-2'-furyl]adenine) in the cAMP assay. Even at very high concentrations (subtoxic), we could not obtain any consistent inhibition of adenylyl cyclase activity, stimulated with orexin or forskolin (an exogenous cell-permeable adenylyl cyclase activator), with any of the inhibitors, and did thus not use these in the ERK assay (data not shown). We then, instead, investigated whether endo- and exogenous phosphodiesterases could be used to decrease any putative cAMP elevation. cAMP production was assayed in the presence and absence of the phosphodiesterase inhibitor IBMX (3-isobutyl-1-methylxanthine) and even with transient expression of the cAMP-specific phosphodiesterases PDE4D1 and -D3 (38) to further reduce any putative cAMP elevations (39). PDE4D3 is found both in the cytosol and in the plasma membrane (40), whereby it may be even more effective in removing cAMP close to the generation sites. Exclusion of IBMX attenuated most of the ß2-adrenoceptor response whereas the forskolin response was more resistant. Successively stronger inhibition was obtained with coexpressed phosphodiesterases, and PDE4D3 inhibited the forskolin response by about 70% (Fig. 9A), which is likely to correspond to the transfection efficacy (forskolin stimulates all the cells whereas the transfected phosphodiesterases are expressed only in a subset of the cells). In contrast, when the cells were transfected with ß2-adrenoceptors together with the PDE4D3, cAMP elevation was entirely blocked (ß2-adrenoceptors are likely to be expressed in the same cells as the phosphodiesterases) (Fig. 9A). Although 100 nM orexin-A used for investigations of ERK phosphorylation gives a saturating ERK response (Fig. 2), it is below the EC50 value for cAMP elevation (16), and thus produces a rather modest cAMP elevation even in the presence of IBMX, but no measurable response in the absence of IBMX in control cells (ctrl) or cells expressing PDE4D1 or -D3 (Fig. 9A). Because the measurements of ERK phosphorylation are conducted in the absence of IBMX with 100 nM orexin-A, it appears unlikely that any significant orexin-A-induced cAMP elevation would occur and affect ERK phosphorylation. However, there could be local cAMP elevations, difficult to detect despite the sensitivity of the cAMP measurements. We thus transfected CHO cells transiently with glutathione-S-transferase (GST)-ERK with or without PDE4D1 and -D3 and evaluated the effect of these phosphodiesterases on ERK phosphorylation. A small inhibition, significant only for PDE4D3, was observed (Fig. 9B). We further evaluated the ability of cAMP to stimulate ERK phosphorylation using the cell-permeable cAMP analog 8-Br-cAMP (500 µM) and forskolin (10 µM) in the ERK assay. Only very low (
10% of the orexin response) and transient activation of ERK phosphorylation was obtained with either of these compounds (Fig. 9C), although a very significant intracellular cAMP elevation was achieved (data not shown). We also applied 8-Br-cAMP and forskolin together with thapsigargin to see whether cAMP and Ca2+ influx could synergize in ERK phosphorylation. 8-Br-cAMP did not affect the thapsigargin response whereas forskolin reduced it (Fig. 9D).
View larger version (29K):
[in this window]
[in a new window]
|
Fig. 9. The Dependence of the OX1-Receptor-Mediated ERK Activation on cAMP in CHO Cells
A, cAMP elevations in cells transiently transfected with ß2-adrenoceptor were measured in response to forskolin (10 µM), isoproterenol (10 nM), or orexin-A (100 nM) in the presence and in the absence of the phosphodiesterase inhibitor IBMX. Some cells were additionally transiently transfected with the cAMP phosphodiesterases PDE4D1 and -D3. The comparison for each orexin-A response is to the corresponding basal level. B, ERK phosphorylation upon stimulation with 100 nM orexin-A in cells transiently transfected with PDE4D1 and -D3. The comparison is to the noninhibited response (100%), indicated by the dotted line. C, pERK levels were measured after 10 and 30 min preincubation with 8-Br-cAMP (500 µM) and forskolin (10 µM). D, Cells were incubated with thapsigargin (1 µM) for 3 and 30 min. Before addition of thapsigargin, the cells were pretreated with 100 µM 8-Br-cAMP for 30 min or with 10 µM forskolin for 3 min. ns, Not significant; isoprot, isoproterenol.
|
|
Ca2+ Influx Regulates OX1 Receptor Signaling to Multiple Effectors
As evidenced in Figs. 3, 5, and 7, OX1 receptor-mediated signaling to ERK seems to obligatorily depend on Ca2+ influx. However, the different efficacy of OX1 receptor to induce Ca2+ influx and ERK phosphorylation (Fig. 2) and the different efficacy and time dependency of OX1 receptor-induced and purely Ca2+-stimulated ERK activation (thapsigargin, ionomycin; Fig. 3) suggest that Ca2+ elevation is not the only factor regulating ERK activity upon OX1 receptor activation. Based on this and previous studies (see Discussion), two reasonable explanations can be envisaged: 1) that Ca2+ influx acts on different effectors (e.g. PLC, PKC, RasGEFs) and stimulates these directly or together with other signals from orexin receptors or 2) that Ca2+ influx acts at a level closer to orexin receptors and enables them to couple to other signals leading to ERK activation. We have previously suggested that the former scheme (scheme 1) could explain the coupling of orexin receptors to PLC at low orexin-A concentrations (11), although no PLC isoform is known to be activated in this manner (i.e. by coactivation via a GPCR signal and Ca2+ but by neither of these alone). We decided to look into a third OX1 receptor-mediated response in CHO cells, cAMP elevation. We have previously shown that OX1 receptors in CHO cells couple to cAMP elevation via high-efficacy PKC and low-efficacy Gs couplings. Because PKC can only stimulate adenylyl cyclase in the presence of simultaneous Gs stimulus, the low-efficacy coupling of the receptors to Gs determines the overall efficacy (16). In agreement with that study, 1 µM orexin-A was able to strongly elevate cAMP in OX1 receptor-expressing CHO cells, a response that was inhibited by reduction of the extracellular Ca2+ (Fig. 10A). Also the Ca2+ channel blocker Ni2+ (5 mM) inhibited the cAMP elevation to a similar degree as it inhibited orexin-induced ERK phosphorylation (data not shown). The direct adenylyl cyclase activator forskolin and cholera toxin (CTx; Gs stimulus) even more potently stimulated adenylyl cyclase activity (Fig. 10B). Ca2+ elevation did not stimulate adenylyl cyclase in CHO cells under basal conditions or under simultaneous Gs (CTx) or forskolin stimulation (Fig. 10C), confirming that there is no Ca2+-stimulated adenylyl cyclase activity in CHO cells (see also Ref.16). Thapsigargin gave a very weak stimulation in the presence of TPA (Fig. 10C), a response that is dependent on conventional PKC(s) (16). Consequently, and in agreement with a previous report (16), OX1 receptor-induced Ca2+ elevation (influx) cannot be directly stimulatory to adenylyl cyclase, but the effect of Ca2+ influx elevation must occur at a level upstream of adenylyl cyclase. In the light of similar results with respect to ERK phosphorylation and PLC activation (see Introduction and Discussion), we conclude that this is likely to reflect a general process in which intracellular Ca2+ is acting proximally to the orexin receptor, being a general obligatory factor for multiple signal coupling, and not directly on different downstream effectors (see Discussion).
View larger version (18K):
[in this window]
[in a new window]
|
Fig. 10. cAMP Response to OX1 Receptor Activation and Other Stimuli in CHO Cells
A, cAMP elevation in response to 1 µM orexin-A in cells in 1 mM (normal TBM), 1 µM and 140 nM extracellular Ca2+. Stimulation of basal adenylyl cyclase activity with PKC (TPA), forskolin, and Gs (CTx). TPA and forskolin were added to the cells for the time of the experiment (10 min), whereas in the case of CTx, the cells were pretreated for 18 h. C, Ca2+ sensitivity of adenylyl cyclase in CHO cells. The cells treated as in B were additionally stimulated with 1 µM thapsigargin or 1 µM ionomycin. For the sake of clarity, the data are normalized to the basal level in each treatment type. ctrl, Control.
|
|
Other Influx Pathways Can Substitute for the Receptor-Operated Pathway in Supporting ERK Phosphorylation via OX1 Receptors
The data in Fig. 5 suggest that inhibition of receptor- and store-operated Ca2+ influx pathways (Ni2+) attenuates orexin response whereas inhibition of the store-operated pathway alone (2-APB) does not. As orexin receptors activate both pathways, either both, in a redundant manner, or the receptor-operated pathway alone may support ERK phosphorylation. Lacking a suitable specific inhibitor for the receptor-operated pathway alone, we reasoned that thapsigargin, which activates the store-operated Ca2+ channels, but simultaneously inhibits the orexin receptor-operated influx (11), could be used to resolve this matter. To confirm our previous results obtained for Ca2+, we resorted to fura-2 measurement of the entry of other divalent cations. Ba2+ and Sr2+ appeared to selectively enter through the orexin receptor-operated channel but not the store-operated channels of CHO cells (Fig. 11, A and B; only shown for Ba2+). Application of thapsigargin before orexin-A fully blocked the orexin-A-induced Ba2+ entry (Fig. 11C). This confirms that, indeed, thapsigargin attenuates orexin receptor-induced (receptor-operated) influx by some mechanism, possibly by direct inhibition (41). Similarly, the activity of store-operated Ca2+ channels has been observed to inhibit the activity of other Ca2+ influx pathways (42, 43).
View larger version (8K):
[in this window]
[in a new window]
|
Fig. 11. Orexin-A-Induced Ba2+ Influx in OX1 Receptor-Expressing CHO Cells Is Blocked by Thapsigargin
The cells were held in nominally Ca2+-free TBM ([Ca2+] 1 µM), and 100 nM orexin-A, 1 µM thapsigargin, 2 mM Ba2+, or 1 mM Ca2+ were applied by perfusion. A and B, Orexin-A, but not thapsigargin, induces Ba2+ influx. C, Treatment of the cells with thapsigargin fully attenuates the response to subsequent application of orexin-A. The responses were measured using microfluorometric Ca2+ imaging; the traces represent averages of about 20 cells. Standard error is shown for every 10th point. s, Second.
|
|
We thus treated CHO cells with thapsigargin for 60 min, to make sure that the thapsigargin-stimulated ERK activity would be nullified, followed by stimulation with orexin-A, and performed measurements of ERK phosphorylation (Fig. 12, A and B). As already shown above, thapsigargin caused a transient ERK phosphorylation (Fig. 12A, "thaps 10'"), which had fully decayed at 60 min (Fig. 12A, "thaps 60'"). Addition of orexin-A after thapsigargin (Fig. 12A, "thaps 60'oxA 10'") led to ERK phosphorylation of a similar magnitude as the control response (Fig. 12A, "oxA 10'"). Thus, although thapsigargin fully blocks the subsequent OX1 receptor-mediated Ca2+/Ba2+ elevation (Fig. 11C; see also Ref.11), it does not inhibit ERK phosphorylation via OX1 receptors. In another experiment, we treated the cells with thapsigargin for 60 min whereafter extracellular [Ca2+] was reduced to 140 nM for 1 min with EGTA. Subsequent addition of 100 nM orexin-A (Fig. 12B, "Thaps 60'EGTA 1'oxA 10'") or restoration of extracellular [Ca2+] of 1 mM (Fig. 12B, "Thaps 60'EGTA 1'Ca2+ 10'") did not increase ERK phosphorylation; however, simultaneous stimulation with both Ca2+ and orexin-A elevated pERK to the same level as in control cells (Fig. 12B, "Thaps 60'EGTA 1'oxA+Ca2+ 10'"). These experiments strongly suggest that Ca2+ influx via the store-operated pathway can fully support orexin receptor signaling to ERK. In a similar manner, cAMP elevation via OX1 receptors was not inhibited by thapsigargin pretreatment (Fig. 12C), supporting this conclusion.
View larger version (27K):
[in this window]
[in a new window]
|
Fig. 12. The effect of Thapsigargin Pretreatment on Orexin-Induced ERK (A, B, and D) and cAMP (C) Responses
The concentration of orexin-A was 100 nM (1 µM in C), thapsigargin (thaps), 1 µM, 2-APB, 100 µM, and EGTA, 0.9 mM. A, ERK response to thapsigargin and orexin-A in control cells (thaps 10' and oxA 10', respectively) and in cells pretreated with thapsigargin for 60 min (thaps 60' and thaps 60'oxA 10', respectively). B, ERK responses after 60-min thapsigargin pretreatment and EGTA (1 min). The responses are to orexin-A (thaps 60'EGTA 1'oxA 10'), restoration of extra- (and intra-) cellular Ca2+ (thaps 60'EGTA 1'Ca2+ 10') and both together (thaps 60'EGTA 1'oxA+Ca2+ 10'). There was no significant difference between the orexin responses (oxA 10', thaps 60'oxA 10' and thaps 60'EGTA 1'oxA+Ca2+ 10'). C, cAMP elevation in response to orexin-A in control and thapsigargin-pretreated (1 µM, 10 min) cells in 1 mM extracellular Ca2+. D, Effect of 2-APB on orexin-A response after 60-min thapsigargin pretreatment. The cells were treated with thapsigargin for 60 min, after which orexin-A was added for 10 more min. 2-APB was added 5, 25, or 55 min before orexin-A. ', Minutes; thaps, thapsigargin; oxA, orexin-A.
|
|
If the store-operated pathway is indeed important for the thapsigargin-mediated "priming" of orexin-A signaling, this effect should be reversible upon blocking of the store-operated pathway. CHO cells were treated with thapsigargin for 60 min before applying orexin-A, and 2-APB was applied at different times before orexin-A. Upon incubation of the cells with 2-APB for 5 min before orexin-A, a slight inhibition of the pERK response to orexin-A was seen, and the response was gradually reversed in cells treated with 2-APB for 25 and 55 min (Fig. 12D). This confirms the necessity of store-operated Ca2+ influx for OX1 receptor signaling to ERK in thapsigargin-treated cells.
|
DISCUSSION
|
|---|
We have in this study demonstrated phosphorylation of ERK1/2, indicative of activation of the ERK signaling pathway, by orexin peptides. Upon recombinant expression, both OX1 and OX2 receptors shared this signaling pathway, and the response was also seen in native striatal neurons. For practical reasons, we chose to perform more extensive studies on recombinant OX1 receptors in CHO cells. Further experiments showed that the relative potencies of orexin peptides, orexin-A and orexin-B, were the same as seen in Ca2+ measurements (20). By the use of an extensive panel of inhibitors, multiple intracellular pathways, including Ras, Src, PI3K, and PKC, are suggested to be involved in this response. In accordance with other studies (reviewed in Refs.17 and 18), this shows that there is considerable diversity in the pathways leading from GPCRs to ERK activation. The results also point at a central role of the OX1 receptor-operated Ca2+ channel(s), but not Ca2+ release, in the signal to ERK and also for the overall signaling of OX1 receptors.
One intriguing aspect of the orexin-mediated ERK phosphorylation was the temporal profile of the response, i.e. a very fast activation and a prolonged response. This was very different from the ERK response caused by Ca2+ elevation with thapsigargin and ionomycin, and from the clearly more transient ERK response observed for OX1 receptors in a previous report (19). Whereas very brief ERK activation may not affect the cell cycle, a longer-lasting activation may lead to proliferation, cell cycle arrest, or differentiation (reviewed in Refs.44 and 45). In this study we have not measured ERK phosphorylation at any time point longer than 3 h, but our other data suggest that OX1 receptor-mediated ERK activation is protective against apoptotic cell death in the long run (65), and further experiments to explore the long-term effects of ERK activation in our system are underway.
We have previously shown that OX1 receptors expressed in CHO cells activate a Ca2+ influx pathway as the primary response (11, 21). Recently, we have isolated this pathway electrophysiologically (12). At low orexin concentrations, stimulation of PLC by OX1 receptors in CHO cells is apparently obligatorily dependent on Ca2+ influx (11, 21), although Ca2+ itself is only a weak stimulant of PLC activity in these cells (11). This has been difficult to explain based on known PLC regulation. In the present study, we have observed and investigated the similarly central role of the Ca2+ influx pathway in ERK activation. In OX1 receptor signaling, Ca2+ could act either at the level of receptor-transducer or at the level of an effector further downstream, as assumed previously (11) and shown for many targets, such as RasGEFs and Ras-GTPase-activating proteins and conventional PKCs (29, 46). Although these two possibilities are nearly impossible to separate experimentally with respect to ERK activation (several Ca2+-dependent targets are, in any case, likely to be part of the signal cascade), the indirect evidence from the ERK assay and direct evidence from other assays point at the former model, i.e. that Ca2+ is most importantly acting upstream of different effectors. Specifically, 1) orexin concentration-response curves with respect to ERK activation are markedly (10-fold) right-shifted as compared with the Ca2+ response. 2) Ca2+ elevation alone is a weaker and much more transient stimulus for ERK activation than OX1 receptor activation, similar to the case of PLC (see above and Ref.11). 3) Thapsigargin attenuates orexin receptor-mediated Ca2+ elevation but not ERK phosphorylation. These three arguments show that Ca2+ influx cannot be the only signal to ERK from orexin receptors but Ca2+ could still act synergistically with other OX1 receptor signals at the effector level. However, this becomes less likely in light of the evidence obtained from the signaling of orexin receptors to other effectors in CHO cells, adenylyl cyclase, and PLC. 4) Adenylyl cyclase activation by OX1 receptors can be blocked by the removal of extracellular Ca2+, although there clearly is no significant Ca2+-stimulated adenylyl cyclase activity present in our CHO cells. Based on this evidence we consider it likely that (locally) elevated Ca2+ level allows OX1 receptors to couple to other intracellular signal cascades in CHO cells. This Ca2+ elevation can be obtained via activation of receptor-operated Ca2+ channels by OX1 receptors themselves, but also via other Ca2+ channels, at least the store-operated Ca2+ channels, as demonstrated here. This hypothesis also offers a natural explanation for the activation of PLC via OX1 receptors at low orexin concentrations (see above), an explanation that is in agreement with the general knowledge on PLC regulation. Thus, even PLC data support this model (point 5). However, it is possible that some responses arise in part from binding of Ca2+ to downstream effectors. It is also possible that Ca2+ dependence of orexin receptors’ signal coupling may be influenced by the cellular background and the type of signal. Identification of the direct interaction partners of OX1 receptors would shed light on this.
Our results suggest that conventional and atypical classes of PKC, PI3K, Ras, and Src cooperate in the orexin-mediated ERK activation. For each of these mediators, we have used several inhibitors to obtain as reliable results as possible. We also performed Ca2+ measurements with the inhibitors to ensure that they did not have nonspecific effects on cell/OX1 receptor signaling. In the lack of any effect of the compounds and constructs on Ca2+ signaling, we consider that the effects on OX1 receptor-mediated ERK phosphorylation are specific. We have focused on the measurements of ERK phosphorylation, and thus the internal relationships of the signal cascades cannot be outlined with certainty. However, with knowledge of the recognized signal cascades a crude hypothesis can be constructed. Ras, being very central to the ERK phosphorylation, may shunt the signals directly to Raf and to PI3K (also responsive to other signals such as phosphotyrosines), the latter of which could also lead to phosphorylation of Raf via PDK1 and PKCs (reviewed in Ref.47). Thus, conventional PKCs could lie downstream of PI3K, as suggested by our data, but additional stimulus to PKC could be obtained via diacylglycerol (DAG) and Ca2+, or as applies to atypical PKCs, by Ras. Ras itself could be activated via GEFs activated, for instance, by phosphotyrosines and DAG (reviewed in Refs.29 and 30). Whether there is a small Ras-independent component in the signal cascades leading to ERK is difficult to judge. If there was, this could be mediated, for instance, by PKCs, B-Raf, or PAKs (p21-activated kinases). The role of Ca2+ in the downstream signal cascades is difficult to evaluate because it is so elementary to the orexin receptor signaling at putatively higher level, but Ca2+ influx could act, for instance, on Ras- (and Rap-)GEFs, conventional PKCs, and PYK2. cAMP, via protein kinase A, has previously been suggested to be a strong stimulant of ERK phosphorylation in CHO cells (e.g. see Ref.37). However, in our CHO cells, cAMP alone or together with elevated Ca2+ stimulated ERK phosphorylation only very weakly and transiently. As for the OX1 receptor signaling, our own results are somewhat contradictory: we observed a small (25%) inhibition of the orexin-A response upon overexpression of PDE4D3, yet we could not measure any cAMP elevation even in the absence of the exogenous phosphodiesterases. One possibility is that the exogenous phosphodiesterases interfere with the signal cascades or with cell maturation in a nonspecific manner. Another possibility is that the inhibition is specific and PDE4D3 removes some very local cAMP hotspots that cannot be detected in cAMP measurements. Even if the latter were true, cAMP signaling would not play a major part in OX1 receptor-stimulated ERK phosphorylation.
In conclusion, our results suggest that OX1 orexin receptor coupling to Ca2+ channel activation lies upstream of their coupling to many other signal pathways, including MAPK pathways, adenylyl cyclase, and PLC, a model radically diverging from the conventional GPCR signaling. We have shown that even in the absence of the activation of the OX1 receptor-operated pathway, the store-operated pathway can function in its place, an effect that can be fully reversed upon specific inhibition of the store-operated pathways with 2-APB. Somewhat puzzling is the much slower effect of 2-APB than EGTA in blocking the ERK response (Fig. 12, B and D). A likely explanation to this is found in the different properties of EGTA and 2-APB. First, EGTA is a Ca2+ chelator and, although membrane impermeant, can extract Ca2+ even from the intracellular side of the plasma membrane. Thus, EGTA may reduce the intracellular Ca2+ concentration much faster to a level that is below that required for "priming" of OX1 receptor signaling. Second, whereas EGTA effectively hinders all Ca2+ influx, 2-APB may not block all the store-operated influx pathways to the cell (26), and some remaining Ca2+ influx may be able to sustain, although not alone support, the primed orexin receptor response.
The question remains whether orexin receptors always couple to some Ca2+ channels themselves, as suggested by the data from neurons, endocrine cells, and recombinant systems (Refs.9 and 48 ; reviewed in Ref.3), or whether they might act as coincidence detectors utilizing other Ca2+ elevations for their signal coupling. It is also unknown how the Ca2+ sensitivity of the signaling is brought about. Because orexin receptors apparently do not possess any Ca2+-binding motif themselves, they may have to interact with some intracellular Ca2+-binding protein. An interesting possibility is that some of the recently isolated Ca2+-binding proteins putatively interacting with orexin receptors might serve this function (Holmqvist et al., unpublished), and, consequently, we will concentrate future efforts in this direction. The results also indicate marked ERK signaling upon orexin receptor stimulation of cultured CNS neurons. The mechanisms and the identity of the responding cells will be other interesting questions to be addressed in future studies.
|
MATERIALS AND METHODS
|
|---|
Cell Culture
CHO-OX1 and -OX2 cells, expressing OX1 and OX2 orexin receptors, respectively, and neuro-2a-hOX1 cells have been described previously (11, 10, 20). CHO cells were grown in Ham’s F-12 medium (GIBCO, Paisley, UK) and neuro-2a cells were grown in DMEM, each supplemented with 100 U/ml penicillin G (Sigma Chemical Co., St. Louis, MO), 80 U/ml streptomycin (Sigma), 400–500 µg/ml geneticin (G418; GIBCO), and 10% (vol/vol) fetal calf serum (GIBCO) at 37 C in 5% CO2 in an air-ventilated humidified incubator in 260-ml plastic culture flasks (75-cm2 bottom area; Greiner Bio-One GmbH, Frickenhausen, Germany). For microfluorometry and Ca2+ measurements, the cells were grown on uncoated circular glass coverslips (diameter, 13 or 25 mm; Menzel-Gläser, Braunschweig, Germany) and for other experiments on circular plastic culture dishes (inner diameter, 32, 52 or 82 mm; Greiner). The confluence of the dishes, independent of the assay or whether the cells were transfected or not, was about 70% by the time of the experiments.
Primary cultures of striatal neurons were prepared from newborn (d 0) Sprague Dawley rats (Scanbur-BK; Sollentuna, Sweden; permission no. C 105/4, Tierps Tingsrätt, Tierp, Sweden, 2004) as previously described for posterior hypothalamic neurons (49). Briefly, striati were dissected from coronal slices. After trypsinization and wash, the cells were dissociated by mild trituration in nutrient medium [Minimum Essential Medium (GIBCO) supplemented with 10% fetal calf serum; 8 g/liter glucose; 2 mM glutamine; 0.1 U/ml insulin; 2 g/liter NaHCO3; and 10 mM HEPES, pH 7.4]. Washed cells were resuspended in the nutrient medium and plated on polyethylenimide-coated glass coverslips (diameter, 13 mm). The cells were cultured in this medium overnight in the incubator as above. The following day, one volume of Neurobasal medium (GIBCO) supplemented with 2% B27 (GIBCO) and 35 µg/ml gentamicin (GIBCO) was added on top of the nutrient medium. On d 5, 10 µM cytosine arabinoside (Sigma) was added to inhibit proliferation of fibroblasts and glia. The cells were studied after 7–14 d in culture.
Chemicals
2-APB, CTx, GF109203X, Gö6976, ionomycin, LY294002, SQ 22536, SU6656, and U0126 were from Calbiochem (La Jolla, CA) and aprotinin, 8-Br-cAMP, 2',3'-dideoxyadenosine, EGTA, forskolin, leupeptin, MDL-12,330A, Na+-pyrophosphate, Na+-orthovanadate, NaF, phenylmethylsulfonyl fluoride, p-nitrophenol phosphate, probenecid [p-(dipropylsulfamoyl)benzoic acid], TPA and Triton X-100 (t-octylphenoxypolyethoxyethanol) were from Sigma. Human orexin-A and -B were from Neosystem (Strasbourg, France), and fura-2 acetoxymethyl ester was from Molecular Probes, Inc. (Eugene, OR). U-73122 and wortmannin were from Tocris Cookson Ltd. (Bristol, UK), thapsigargin was from RBI (Natick, MA), and glycerol and NiCl2 were from Merck AG (Darmstadt, Germany). EDTA was from Boehringer Mannheim GmbH (Mannheim, Germany) and [3H]adenine, [14C]cAMP, [3H]inositol, and Tween20 (polyoxyethylene sorbitan monolaurate) were from Amersham Biosciences (Buckinghamshire, UK). SB-334867 (50) was a generous gift from Dr. Neil Upton (Neurology CEDD, GlaxoSmithKline Pharmaceuticals, Harlow, UK).
Media
TES-buffered medium (TBM) consisted of 137 mM NaCl, 5 mM KCl, 1 mM CaCl2, 1.2 mM MgCl2, 0.44 mM KH2PO4, 4.2 mM NaHCO3, 10 mM glucose, and 20 mM TES (2-[(2-hydroxy-1,1-bis[hydroxymethyl]ethyl)amino] ethane sulfonic acid) adjusted to pH 7.4 with NaOH. Lysis buffer was composed of 50 mM HEPES and 150 mM NaCl (pH 7.5) supplemented with 10% (vol/vol) glycerol, 1% (vol/vol) Triton X-100, 1.5 mM MgCl2, 1 mM EDTA, 10 mM Na+-pyrophosphate, 1 mM Na+-orthovanadate, 10 mM NaF, 250 µM p-nitrophenol phosphate, 10 µg/ml aprotinin, 10 µg/ml leupeptin, and 1 mM phenylmethylsulfonyl fluoride. Laemmli sample buffer was composed of 50 mM Tris-HCl (pH 6.8) supplemented with 1 mM dithiothreitol, 2% sodium dodecyl sulfate (SDS) (wt/vol), 10% glycerol (vol/vol), and 0.1% bromophenol blue (wt/vol).
Expression Vectors
Most of the plasmid constructs used in this study were generous gifts from other scientists, whom we gratefully acknowledge. pEBG ERK2 (GST-ERK2) was from Dr. Bruce J. Mayer (University of Connecticut Health Center, Farmington, CT), pMCL MKK1 Kd [dominant-negative (kinase-dead; K97M) MKK1] (51) from Dr. Natalie G. Ahn (University of Colorado, Boulder, CO), pEGFP-C1-PLC-PH (fusion of GFP and PH domain of PLC1) (52) from Dr. Tobias Meyer (Stanford University School of Medicine, Stanford, CA), pEGFP-C1-PKCwt (fusion of GFP and PKC) (53) from Dr. Johanna Ivaska (VTT Medical Biotechnology, Centre for Biotechnology, Turku, Finland), FLAG-PKC (T410A)/pCMV5 [dominant-negative (phosphorylation site mutant) PKC] and FLAG-PKC (KW)/pCMV5 [dominant-negative (ATP-binding site mutant) PKC] (36) from Dr. Maggie M. Chou (University of Pennsylvania School of Medicine, Philadelphia, PA), pcDNA3-GFP-PTEN; A4 (S380A, T382A, T383A, S385A-GFP-PTEN, more active than wild-type PTEN) (54) was from Dr. William R. Sellers (Dana-Farber Cancer Institute, Boston, MA), pSG5 p85 iSH2-N [dominant-negative (non-p110-binding) p85] (55) from Dr. Michael D. Waterfield (Ludwig Institute for Cancer Research, London, UK), pCMV5 RF Src [dominant-negative (K295R, Y527F) Src] (56) was from Dr. Joan S. Brugge (Harvard Medical School, Boston, MA), pSR-CSK [wild-type Csk (C-terminal Src kinase)] (57) from Dr. Masato Okada, pEXV3 H-Ras N17 [dominant-negative H-Ras (S17N)] (58) from Dr. Sally J. Leevers (Cancer Research UK London Research Institute, London, UK), pCEFLHA-K-Ras-N17 and pCEFLHA-N-Ras17 [dominant-negative K-, respectively, N-Ras (S17N)] (34) from Dr. Piero Crespo (Biomédicas, Consejo Superior de Investigaciones Centíficas, Madrid, Spain), pRK5-Rap1A S17N [dominant-negative (S17N) Rap1A] from Dr. Jean de Gunzburg (Institut Curie-Section de Recherche, Paris, France), pCAG Ral S28N [dominant-negative (S28N) RalA] (59) from Dr. Larry A. Feig (Tufts University School of Medicine, Boston, MA) and pCMV5 PDE4D1 and pcDNA3.1-PDE4D3 (cAMP phosphodiesterases PDE4D1 and -D3) (60) from Drs. Marco Conti and Wito Richter (Stanford University School of Medicine, Stanford, CA). Human ß2-adrenoceptor in pcDNA 3.1 was from Guthrie cDNA Resource Center (http://www.cdna.org) and pEGFP-C1, used to identify transfected cells in some experiments, was from CLONTECH (Palo Alto, CA).
Transfection
CHO-OX1 cells were grown on 52-mm (inner diameter) plastic culture dishes to 40–50% confluence. The dishes were washed with PBS, and OPTI-MEM (GIBCO, Paisley, UK) was added. The cells were transfected using Lipofectamine reagent (Invitrogen Corp., Carlsbad, CA). After 5 h this medium was replaced with fresh Ham’s F-12 medium with all the usual supplements (see above). The culture medium was replaced 19 h later with serum-free medium, and the cells were starved for 24 h before stimulation (see below). Transfection efficiency was 40–70% as determined using expression of GFP and some functional criteria (e.g. phosphodiesterase activity). Transfection of the cells was performed to specifically inhibit selected pathways toward ERK phosphorylation or cAMP generation or to introduce probes for measurement of PLC activity (this was performed with cells on glass coverslips). For the former purpose, GST-ERK (pEBG ERK2) was introduced together with the inhibitor plasmid to circumvent the problem of less than 100% transfection efficiency. The much higher molecular weight of the fusion protein allows separation from the endogenous ERK an
TOP
ABSTRACT
INTRODUCTION
