Cross-talk between 5-Hydroxytryptamine Receptors in a Serotonergic Cell Line

The study of signaling cascades and of functional interactions between 5-hydroxytryptamine (5-HT) receptor pathways with heterogenous brain cell populations remains an arduous task. We took advantage of a serotonergic cell line to elucidate cross-talks between 5-HT receptors and to demonstrate the involvement of two 5-HT2 receptor subtypes in the regulation of 5-HT1B/1D function. The inducible 1C11 cell line has the unique property of acquiring within 4 days a complete serotonergic phenotype (1C11* cells), including three 5-HT receptors. 5-HT1B/1D and 5-HT2B receptors are expressed since day 2 of the serotonergic differentiation while 5-HT2A receptors are induced at day 4. We first established that 5-HT2B receptors are coupled with the phospholipase A2 (PLA2)-mediated release of arachidonic acid (AA) and that the activation of 5-HT2B receptors in 1C11*d2 cells inhibits the 5-HT1B/1D receptor function via a cyclooxygenase-dependent AA metabolite. At day 4, this 5-HT2B-mediated inhibition of the 5-HT1B/1D function can be blocked upon concomitant 5-HT2A activation although a 5-HT2A/PLA2 positive coupling was evidenced. This suggests the existence in 1C11*d4 cells of pathway(s) for 5-HT2A receptors, distinct from PLC and PLA2. Finally, this study reveals the antagonistic roles of 5-HT2A and 5-HT2B receptors in regulating the function of 5-HT1B/1D, a receptor involved in neuropsychiatric disorders and migraine pathogenesis.

In view of the pivotal role of serotonin (5-hydroxytryptamine (5-HT)) 1 in neurologic and neuropsychiatric disorders, to identify the mechanisms that mediate the cellular functions controlled by 5-HT is a major challenge. To achieve this goal, the study of signaling cascades and of functional interactions between 5-HT receptor pathways with heterogenous brain cell populations still appears impractical. An alternative is to use a clonal cell line, such as 1C11, expressing a definite set of 5-HT receptors (1).
The 1C11 clone was derived from F9 multipotential embryonal carcinoma cells through immortalization and differentiation (2,3). The 1C11 clone has the properties of a neuroectodermal progenitor able to differentiate into 5-hydroxytryptaminergic cells (1C11*) by induction by dibutyryl cyclic AMP (Bt2cAMP) and cyclohexane carboxylic acid (CCA). The switch from the undifferentiated committed 1C11 cell type to the 1C11* serotonergic cells occurs within 4 days in more than 90% of the cell population.
We recently demonstrated that, in addition to the onset of 5-HT metabolism, storage, and transport, 1C11* cells also acquire 5-HT1B/1D, 5-HT2A, and 5-HT2B receptors (1). Binding and transductional experiments excluded the functional presence of any other 5-HT receptor subtypes. On day 2 of the serotonergic differentiation, 5-HT1B/1D and 5-HT2B receptors became expressed and remained functional until at least day 4, at which time the 5-HT2A receptor was induced. The appearance of the latter receptor coincided with the onset of an active 5-HT transport system. This sequence of events in the inducible 1C11 cell line offers the possibility of exploring cross-talks between the signaling pathways of these three G-protein coupled receptors within a complete serotonergic phenotype.
Many studies have established that 5-HT1B/1D receptors are negatively coupled with adenylate cyclase through Gi (GTP-binding protein) (4). Acting as terminal or somatodendritic autoreceptors, they mediate and modulate serotonergic neurotransmission (5). They can also function as terminal heteroreceptors by inhibiting, after agonist stimulation, the release of other neurotransmitters such as acetylcholine, glutamate, dopamine, norepinephrine, and ␥-aminobutyric acid (6 -8). Thus, signals interfering with the G i -coupled 5-HT1B/1D system may act widely on neuronal responses. In addition, 5-HT1B/1D receptors are involved in cardiovascular functions, vasospasm, and migraine (9). Because of both the pivotal role of 5-HT1B/1D receptors in neurotransmission and the clinical impact of 5-HT1B/1D receptors agonists like sumatriptan (10) in the acute treatment of migraine, studying the regulatory aspects of 5-HT1B/1D receptor function is of great importance.
Two distinct receptors of the 5-HT2 class are also induced during the serotonergic differentiation of 1C11 cells. 5-HT2A and 5-HT2B receptors display a strong amino acid sequence homology and share close but distinct pharmacological profiles (11). Each of these receptors have been described as mediating phosphatidylinositol (PI) hydrolysis through activation of phospholipase C (PLC␤). The 1,2-diacylglycerol (DAG) thus obtained activates PKC (12), whereas inositol 1,4,5-triphosphate (IP3) increases intracellular Ca 2ϩ (13). Nevertheless, the 5-HT2B receptor fails to significantly stimulate PI hydrolysis in rat stomach (14) as well as in rat vasculature (15). This suggests that signaling processes coupled to 5-HT2B receptors, but distinct from PLC, may also occur in vivo. Accordingly, Launay et al. (16) have reported a 5-HT2B-dependent activation of the p21 ras /MAP kinase signaling cascade. In addition, 5-HT2 receptor subtypes are found in regions of the hippocampus and cerebral cortex where 5-HT stimulates the neuronal secretion of arachidonic acid (AA) (17,18). Both 5-HT2A and 5-HT2C, which is a third member of the 5-HT2 receptor family, are likely to be involved in this release because the stable transfection of CHO cell lines with the corresponding human cDNAs activates the phospholipase A2 (PLA2)-mediated AA release (19,20). The PLA2/AA pathway could also be sensitive to 5-HT2B as indicated by the colocalization of this receptor with the 5-HT2A and 5-HT2C ones in the hippocampus (21)(22)(23).

Materials-Forskolin
Measurement of 5-HT1B/1D Receptor Response-Cells were washed twice with HBSS (with calcium and magnesium) containing 10 mM HEPES (pH 7.4) and then preincubated in 500 l of 10 mM HEPES (pH 7.4) (wash buffer) per well for 15 min in a CO 2 incubator (5% at 37°C). Where indicated, inhibitors were added during this preincubation period. 5-HT1B/1D receptor-mediated response was followed by measuring (15 min at 37°C) the 5-CT-induced inhibition of cAMP accumulated in the presence of 1 M forskolin and 0.1 mM rolipram, a phosphodiesterase inhibitor. Cellular cAMP content was measured by radioimmunoassay (1) and normalized to protein content as determined by the bicinchoninic acid assay (Pierce, Chichester, UK). For each experiment, data were expressed according to the response obtained with 1 M forskolin (100%).
IP Accumulation Measurements-Cells were labeled with 1 Ci/ml myo-[ 3 H]inositol in serum-free medium for 24 h at 37°C. Total IP (inositol monophosphate, inositol bisphosphate, and IP3 are collectively referred to as IP) accumulation in response to agonist stimulation in the presence of 20 mM LiCl for 10 min at 37°C was determined as described previously (25). Radiolabeled IP was separated according to the ion exchange method of Berridge et al. (26).
Arachidonic Acid Release-Cells were labeled with 0.1 Ci/ml [ 14 C] AA for 4 h at 37°C (5% CO 2 ). Under these conditions, more than 90% of total radioactivity was taken up by the cells. After labeling, cells were washed three times with HBSS containing calcium and magnesium supplemented with 20 mM HEPES and 0.1% BSA (bovine serum albumin) (experimental medium). For calcium-free experiments, cells were washed with the same experimental medium without calcium. Between two washes, cells were incubated for 5 min at 37°C. After the last wash, cells in 1 ml of experimental medium were exposed to the indicated drugs, and aliquots (100 l) were removed after 10 min for [ 14 C] measurement.
Data Analysis-Concentration-response data were fitted by non-linear regression to the model: R ϭ (R max )/[1 ϩ (EC 50 /A) n ], where E is the measured response at the given agonist concentration A, R max is maximal response, EC 50 is the concentration of agonist producing halfmaximal response, and n is the slope index. The non-parametric Kolmogorov-Smirnov test (KS test) was used for statistical comparisons.
To determine whether the observed [ 14 C]AA release originated from activation of PLA2 or derived from the PLC/DAG lipase pathway, 14 C-labeled 1C11*d2 cells were exposed for 15 min to 100 M mepacrine, a PLA2 inhibitor, prior to the addition of agonists. As shown in Fig. 1B, this completely inhibited the effect of up to 10 M 5-HT on the release of [ 14 C] radioactivity. The involvement in [ 14 C] release of 5-HT2B receptors via PLA2 coupling could be further evinced by using DOI treatment (1 M). DOI-mediated AA release was also fully blocked by mepacrine (Fig. lB). Mepacrine, however, had no significant effect on DOI-mediated PI hydrolysis. Cellular IP accumulation measured after DOI treatment in the presence of 100 M mepacrine, amounted to 124 Ϯ 18% above the basal level (mean Ϯ S.E., n ϭ 5, p Ͼ 0.05). Accumulation was very similar in the absence of mepacrine (117 Ϯ 16%). Mepacrine alone had no effect on IP accumulation, as well as on basal AA release.
Unlike phospholipase C or phospholipase D, cytosolic PLA2 requires extracellular calcium for its activity (29). Accordingly, when calcium was removed from the incubation medium of 1C11*d2 cells, 5-HT-mediated AA release was reduced to 11 Ϯ 7% above the basal level (mean Ϯ S.E., n ϭ 4). We verified that the basal [ 14 C] release level was not sensitive to the deprivation of calcium. All these data indicate that the 5-HT2B receptor of 1C11*d2 cells is coupled with the cytosolic PLA2 activity.
The PLA2-AA Signal Transduction Pathway Mediated by 5-HT2B Activation Inhibits the 5-HT1B/1D Receptor Function-5-HT1B/1D receptors are linked to G i -mediated inhibi-tion of adenylate cyclase activity (11) and become expressed on day 2 of 1C11 cell differentiation (1). 5-CT, as do other 5-HT1B/1D receptor agonists, inhibits forskolin-stimulated cAMP accumulation (FScA). In contrast, the addition of 5-HT2 receptor agonists had no effect on FScA as well as on the basal cAMP cellular level (27). On the other hand, the IP3 content of 1C11* cells does not respond to the addition of 5-HT1B/1D agonists (1). This suggests that the IP3 hydrolysis pathway mediated by 5-HT2B activation does not interfere with the negative coupling of adenylate cyclase with 5-HT1B/1D receptor.
As shown in Fig. 2, 5-CT (5 nM) inhibits FScA by 50%. The extent of inhibition remains similar with higher 5-CT concentrations. Activation of 5-HT2B receptors by 1 M DOI completely abolished the 5-CT-mediated inhibition of FScA (Fig. 2). This behavior could be related to PLA2 since the direct addition of melittin (2.5 g/ml) mimicked the action of DOI on the 5-HT1B/1D receptor signaling pathway (Fig. 2). In the presence of 100 M mepacrine, this effect of DOI was not observed any more. This indicates that the PLA2/AA signaling cascade governs the 5-HT2B-dependent inhibition of the 5-HT1B/1D receptor pathway.
Since multiple pathways including cyclooxygenase, lipoxygenase, cytochrome P-450, and autooxidation (30) may couple AA metabolism with the 5-HT1B/1D receptor response, we investigated the effect of indomethacin, a cyclooxygenase inhibitor. Addition of a 2 M amount of this substance not only blocked the action of DOI but also enhanced the 5-CT-mediated inhibition of FScA (Fig. 2). 1C11*d2 cells were treated with either 100 M mepacrine or 2 M indomethacin prior to the addition of melittin (2.5 g/ml). In both cases, the 5-CT-mediated inhibition of FScA disappeared, as observed in the presence of melittin alone. Thus, by suppressing cyclooxygenase activity, indomethacin can block the inhibition of 5-HT1B/1D function triggered by either 5-HT2B receptor stimulation or the direct activation of PLA2. These data suggest that a cyclooxygenasedependent metabolite of AA sustains the negative regulation of the 5-HT1B/1D response by 5-HT2B receptors.
In 1C11*d4 Cells, the PLA2/AA Signal Transduction Pathway Is Coupled with Both 5-HT2B and 5-HT2A Receptors-Between days 2 and 4 of 1C11* differentiation, the number and the pharmacological profiles of 5-HT1B/1D and 5-HT2B receptors did not vary. On day 4, functional 5-HT2A receptors became induced (1). Thus, with 1C11*d4 cells, it was possible to study the combined effects of these two members of the 5-HT2 receptor class on 5-HT1B/1D receptor function and signaling. Incubation during 10 min of [ 14 C]AA-labeled 1C11*d4 cells with 10 M 5-HT induced a release of [ 14 C] radioactivity equal to 204 Ϯ 19% above the basal release (mean Ϯ S.E., n ϭ 4) (Fig.  3). The released radioactivity is markedly greater than that measured at day 2 (144 Ϯ 12%) (Fig. 1). However, the release could also be fully antagonized by pretreatment with mepacrine (8 Ϯ 9% above basal release; mean Ϯ S.E., n ϭ 6). According to previous reports (18,20), the larger increase in [ 14 C] release observed at day 4 can be partly attributed to an activation of PLA2 by the newly expressed 5-HT2A receptor. A contribution of the PLC pathway may be ruled out by verifying that the pretreatment of 1C11* d4 cells with mepacrine did not change the DOI-mediated PI hydrolysis.
Finally, an involvement in AA release of both 5-HT2 receptor subtypes could be firmly established by using 5-HT2 antagonists with different specificities. LY266070 (1 nM), a selective antagonist of the 5-HT2B receptor (24), reduced by 52% the AA release measured at day 4, whereas it abolished PLA2 activation at day 2 by nearly 100% (Fig. 3). Ketanserin (5 nM), which has only a small effect on the 5-HT2B receptor response at day 2, reduces the [ 14 C]AA release at day 4 by 54% (Fig. 3). These findings are consistent with the preferential 5-HT2A receptor antagonist property of ketanserin, as already described (11).
The Activation of 5-HT2A Receptor Blocks the 5-HT2B-mediated Inhibition of the 5-HT1B/1D Receptor Response-In contrast with the situation at day 2, addition of DOI (1 M) at day 4 had no effect any more on the 5-CT-mediated inhibition of FScA (Fig. 4) (48 Ϯ 5% and 47 Ϯ 4%, in the presence or absence of DOI, respectively; mean Ϯ S.E., n ϭ 4). This behavior suggests that the newly expressed 5-HT2A receptor can antagonize the negative regulation exerted by 5-HT2B on the 5-HT1B/1D function. This was reasserted because the effect of DOI on the 5-CT-mediated inhibition of FScA was restored in the presence of 5 nM ketanserin (not shown).
These results strongly indicate that the 5-HT2A-mediated DOI effect can only be partly accounted for by the PLA2 signaling cascade and that additional transduction pathway(s) must also be involved in the control of 5-HT1B/1D receptor function within the 1C11 serotonergic cell line. DISCUSSION So far, the diversity of 5-HT receptors observed in vivo or in heterogeneous primary cultures has rendered difficult the determination of their precise role in the signaling networks that mediate the 5-HT physiological functions. Consequently, most of the possible functional coupling(s) associated with one given receptor subtype could be defined only by using cDNA-transfected heterologous cell lines. With newly established neuronal cell lines expressing a definite set of receptors, studies may be carried out in conditions closer to in vivo physiological circumstances, particularly concerning active stoichiometries and pathways of G protein coupling. The 1C11 neural-like stem cell is at present a unique in vitro clonal model able to acquire three 5-HT receptors during its serotonergic differentiation program (1). Therefore, this model appears suitable to define the transductional pathways related to the activation of all 5-HT1B/1D, 5-HT2B, and 5-HT2A receptors within the same serotonergic context. The sequential onset of these three receptors along the differentiation process makes it possible to follow the functional interaction(s) between their associated signaling cascades at day 2, when 5-HT1B/1D and 5-HT2B receptors are coexpressed, and at day 4, when 5-HT2A has also been induced.
The in vivo signaling pathways that mediate the physiological functions of 5-HT2B receptors remain largely unknown (14,15). In the present study, we have investigated the PLA2/AA pathway because 5-HT was already known to stimulate the direct release of AA via 5-HT2 receptor subtypes in hippocampal neurons (18,31). As early as day 2 of the 1C11 differentiation, the activation of 5-HT2B receptors with 5-HT or DOI stimulates the release of AA. The PLA2 inhibitor mepacrine completely inhibits this behavior while having no effect on the 5-HT2B-mediated PI hydrolysis (Fig. 1).
Furthermore, the 5-HT2B receptor-mediated AA release observed here may be responsible for the direct activation of the PLA2 activity and does not derive from the phospholipase C-or D-induced PLA2 activation. Indeed, it was totally dependent on extracellular calcium. Thus, inside the 1C11* cells, in addition to its role in activation of PI hydrolysis (27), p21 ras /MAP kinase signaling (16), and NO signaling, 2 the 5-HT2B receptor can also control AA release through the activation of PLA2.
All members of the 5-HT2 receptor family have been detected in those regions of the cerebral cortex and hippocampus where 5-HT stimulates the release of AA (17,18,23). Moreover, after activation of 5-HT2C receptors, PLA2 and cyclooxygenase/lipoxygenase pathways contribute to the formation of cGMP in rat choroid plexus (32). Thus, it is likely that 5-HT2 receptor subtypes mediate in the brain some of the 5-HT physiological functions through the PLA2/AA pathway, but independently of the turnover of phosphoinositides.
On day 2 of 1C11 cell differentiation, the intracellular 5-HT and tryptophan hydroxylase activity becomes measurable. At the same time, 5-HT1B/1D (1500 GTI binding sites/cell) and 5-HT2B receptors (2200 DOI binding sites/cell) are induced (1). The simultaneous presence of these two functional 5-HT receptors renders the 1C11*d2 cells competent to integrate the different signaling inputs mediated by 5-HT. The present study shows that inside 1C11*d2 cells, the 5-HT2B-mediated PLA2 signaling cascade fully represses the negative G i coupling of 5-HT1B/1D receptors with adenylate cyclase (Figs. 2 and 5A). The direct activation of PLA2 by melittin mimics the inhibitory effect exerted by the 5-HT2B receptor on the 5-HT1B/1D-mediated transduction pathway. Mepacrine antagonizes the action of melittin on the 5-HT1B/1D response and prevents the 5-HT2B-mediated inhibition of 5-HT1B/1D signaling. Furthermore, indomethacin, which inhibits the oxidative metabolism of AA through the cyclooxygenase pathway, completely blocks the inhibition of 5-HT1B/1D receptor function triggered by melittin-directed activation of PLA2. Indomethacin also blocks the 5-HT2B receptor inhibitory effect and furthermore potentiates the 5-CT-mediated inhibition of FScA caused by DOI. All these results imply that the actors of the cyclooxygenase metabolism, such as prostaglandins or thromboxanes, intracellularly mediate the interaction between the PLA2/AA and the adenylyl cyclase signaling cascades.
Interestingly, similar potentiating effects of indomethacin have been previously observed in the case of (i) the 5-HT2C receptor-mediated cGMP formation (150% of the control level) 2 J. M. Launay, unpublished data. in the choroid plexus (32), and (ii) the control of 5-HT1B/1D receptor pathway in CHO cells stably transfected by human 5-HT2C receptor cDNA (20). The mechanisms underlying such effects of indomethacin have not yet been elucidated. Regarding 1C11*d2 cells, the enhancement of 5-HT1B/1D function in response to both indomethacin and 5-HT2B activation might reflect an increase of the negative coupling between adenylate cyclase and 5-HT1B/1D. Several signaling pathways could be involved. For instance, when cyclooxygenase activity is blocked by indomethacin, the AA produced by 5-HT2B activation might be converted into a new signal. This signal could activate the 5-HT1B/1D signaling response through another metabolic pathway involving either lipoxygenase, cytochrome P-450 or autooxidation. Accordingly, preliminary experiments with some selective inhibitors of these pathways indicate that the cytochrome P-450 monooxygenase metabolism could be involved. In fact, the addition of either miconazole or SKF-325A, two inhibitors of the cytochrome P-450 system, inhibits the potentiating effect of indomethacin on the 5-HT1B/1D response (not shown). Surprisingly, this potentiating effect could only be observed in the case of 5-HT2B receptor stimulation, but not when PLA2 was directly activated by melittin. Thus, the possibility remains that, by recruitment of additional signaling pathway(s), 5-HT2B receptors also enhance the negative 5-HT1B/1D adenylate cyclase coupling through intracellular Ca 2ϩ mobilization and/or various G protein interactions. Although the effect of indomethacin suggests that the 5-HT1B/1D response is governed by several antagonistic signaling pathways, all our results imply that in 1C11*d2 cells, 5-HT2B receptors exert their dominant negative regulation on the G icoupled 5-HT1B/1D receptor system through cycloxygenase-dependent metabolite(s) of AA.
At day 4 of 1C11 cells differentiation, functional 5-HT2A receptors (400 DOI binding sites/cell) have become expressed (1), and the AA release in response to the addition of 5-HT is greater than at day 2. Because of the different sensitivity of 5-HT2A and 5-HT2B receptors to the antagonists ketanserin and LY266070, we could show that each of these two receptors was coupled with the PLA2/AA pathway (Fig. 3). Despite its coupling with a signaling pathway also recruited by 5-HT2B receptors, 5-HT2A receptor activation prevents the negative regulation exerted by 5-HT2B receptors on the 5-HT1B/1D function (Fig. 5B). The rescue of the 5-HT1B/1D response at day 4 implies that the DOI-induced 5-HT2A activation either blocks or neutralizes the 5-HT2B cyclooxygenase-dependent negative action also triggered by addition of DOI. The coupling of two 5-HT2 receptor subtypes with PLA2 in the 1C11*d4 serotonergic cells, independent of PI hydrolysis, corroborates the few reports dealing with 5-HT-stimulated release of AA, as observed in vivo (18) or in vitro with CHO cell lines stably transfected with human 5-HT2A or 5-HT2C cDNAs (19,20). CHO cells are known to endogenously express 5-HT1B-like receptors. In the 5-HT2A/CHO cell line, 5-HT2A receptor activation has no effect on 5-HT1B signaling. Thus, it is likely that, in 1C11*d4 cells, 5-HT2A receptors can abrogate the 5-HT2B receptor-mediated inhibition of 5-HT1B/1D function by blocking the production of some cyclooxygenase product or its effect, instead of directly acting on the 5-HT1B/1D coupling.
Interestingly, in 1C11*d4 cells, the 5-HT2A-mediated PLA2 signaling cascade only slightly accounts for the blocking action of 5-HT2A receptors. Indeed, PLA2 activation by melittin induces a minor relative decrease (Ϸ16%) of cAMP accumulation, when compared with the complete (Ϸ46%) rescue of the 5-CT mediated inhibition of FScA after DOI treatment (Fig. 4). Consequently, a direct activation of the 5-HT2A receptors is needed to antagonize the negative regulation by 5-HT2B of the 5-HT1B/1D coupling. Therefore, the control of the 5-HT1B/1D function must involve some additional 5-HT2A-mediated transduction pathway(s). These are distinct from the PLC and PLA2 signaling pathways already functional in the 1C11 cell line.
In brain tissue, AA and its metabolites can act as either intracellular or intercellular messengers. Eicosanoids released from cells recognize specific cell receptors (33) and mediate many transcellular actions. They play several roles in neuronal signal transduction (33,34) and seem to be involved in long term potentiation synaptic plasticity, ion channel regulation, neurosecretion, and presynaptic inhibition (30). Moreover, eicosanoids accumulate in the brain during strokes, seizures, or other pathophysiological processes (35,36). Nevertheless, the biological significance of a 5-HT-mediated release of AA in the nervous system is presently unknown. This study has established that in non-transfected 1C11 cells, 5-HT2 receptor-generated eicosanoids support cross-talks between the three 5-HT receptors. Whether, in vivo, a single neuron has each of the 5-HT2A, 5-HT2B, and 5-HT1B/1D receptors is presently unknown. It has been shown, however, that these receptors colocalize in the same brain regions, especially in the hippocampus (22, 23,37). Actually, AA release is coupled to 5-HT2 receptors in this area. It may be that in such places some of the 5-HT2 receptor functions are mediated through the diffusion of the oxidative metabolites of their PLA2/AA coupling. Such metabolites would direct the transduction of signals between cells and modulate neuronal or neurovascular activity.
The main contribution of this study is the demonstration of cross-talks between three 5-HT receptors naturally induced in  Fig. 4. AC, adenylate cyclase; CytP-450m, cytochrome P-450; DHETs, dihydroxyeicosatetraenoic acids; EETs, epoxyeicosatrienoic acids; Lox, lipoxygenase; HPETE, hydroperoxyeicosatetraenoic acids; LTs, leukotrienes; Cox, cyclooxygenase; PGI 2 , prostacyclin; PGs, prostaglandins; TXs, thromboxanes. the course of a serotonergic program. The antagonistic roles of 5-HT2A and 5-HT2B receptors in the regulation of 5-HT1B/1D function are well demonstrated. Links between the regulation of 5-HT1B/1D receptor function and anxiety, depressive or obsessive disorders, and migraine pathogenesis are not yet fully known. It is, however, worth noting that selective 5-HT1B/1D/1F receptor agonists, like sumatriptan or zolmitriptan (10), are effective in the acute treatment of migraine (38). A number of 5-HT2 receptor antagonists are used as prophylactic antimigraine agents, but which 5-HT2 receptor subtype mediates these prophylactic effects is still unclear. A role has been proposed for the 5-HT2C receptor since the 5-HT2A preferential antagonist, ketanserin, is not effective (39). This assumption is further supported by the observation that the 5-HT2C agonist, mCPP, triggers migraine attacks (40). Nevertheless, only trace amounts of 5-HT2C receptor mRNA are detected in meningeal tissues. On the other hand, organ bath experiments have recently suggested that 5-HT2B receptor located on the endothelial cells of meningeal blood vessels would trigger migraine headache, probably via NO formation (41). Thus, 5-HT1B/1D and 5-HT2B receptors are presently the best candidates for involvement in migraine. In this context, a cross-talk analysis between these two 5-HT receptor subtypes as well as 5-HT2A, all present in various meningeal tissues, can be carried out at a cellular level with 1C11* cells.
Finally, the induction of the three 5-HT1B/1D, 5-HT2B, and 5-HT2A receptors is an essential part of the autonomous serotonergic differentiation program of 1C11. These receptors could possibly act as autoreceptors, thus promoting the regulation of the serotonergic phenotype by 5-HT. In fact, the synthesis, metabolism, and storage of 5-HT in 1C11* cells are downregulated by external 5-HT. 3 We now aim at identifying intracellular targets related to the activation of 5-HT receptors and at analyzing the contribution of 5-HT-mediated signals to the control of the serotonergic functions.