Address correspondence and reprint requests to Hubert Vaudry, European Institute for Peptide Research (IFRMP 23), Laboratory of Cellular and Molecular Neuroendocrinology, INSERM U413, UA CNRS, University of Rouen, 76821 Mont-Saint-Aignan, France. E-mail: email@example.com
Endozepines, a family of regulatory peptides related to diazepam-binding inhibitor (DBI), are synthesized and released by astroglial cells. Because rat astrocytes express various subtypes of somatostatin receptors (sst), we have investigated the effect of somatostatin on DBI mRNA level and endozepine secretion in rat astrocytes in secondary culture. Somatostatin reduced in a concentration-dependent manner the level of DBI mRNA in cultured astrocytes. This inhibitory effect was mimicked by the selective sst4 receptor agonist L803-087 but not by the selective sst1, sst2 and sst3 receptor agonists L779-591, L779-976 and L797-778, respectively. Somatostatin was unable to further reduce DBI mRNA level in the presence of the MEK inhibitor U0126. Somatostatin and the sst1, sst2 and sst4 receptor agonists induced a concentration-dependent inhibition of endozepine release. Somatostatin and the sst1, sst2 and sst4 receptor agonists also inhibited cAMP formation dose-dependently. In addition, somatostatin reduced forskolin-induced endozepine release. H89 mimicked the inhibitory effect of somatostatin on endozepine secretion. In contrast the PLC inhibitor U73122, the PKC activator PMA and the PKC inhibitor calphostin C had no effect on somatostatin-induced inhibition of endozepine release. The present data demonstrate that somatostatin reduces DBI mRNA level mainly through activation of sst4 receptors negatively coupled to the MAPK pathway, and inhibits endozepine release through activation of sst1, sst2 and sst4 receptors negatively coupled to the adenylyl cyclase/PKA pathway.
Somatostatin is a multifunctional peptide that is widely distributed in the central and peripheral nervous systems (Epelbaum 1986; Patel et al. 1986). In the brain, somatostatin acting as a neurotransmitter affects cognitive, sensory and autonomic functions (Patel et al. 1995). The effects of somatostatin are mediated through at least five subtypes of receptors (designated sst1–5) which are all negatively coupled to adenylyl cyclase (Csaba and Dournaud 2001). High concentrations of somatostatin binding sites have been found in rat astrocytes both in situ (Mentlein et al. 1990; Krisch 1994) and in culture (Mentlein et al. 1990; Hosli and Hosli 2000). RT–PCR analysis has shown that the mRNAs encoding the sst1, sst2 and sst4 receptor subtypes are actively expressed in astroglial cells (Feindt et al. 1995). It has also been demonstrated that somatostatin inhibits cAMP formation in cultured rat astrocytes (Feindt et al. 1995; Grimaldi et al. 1997), suggesting that somatostatin is an important regulator of glial cell activity.
In the present study we have investigated the effects of somatostatin on DBI mRNA level and endozepine secretion by cultured rat astrocytes and we have determined the signaling pathways involved in the action of somatostatin on astroglial cells.
Materials and methods
Solutions and reagents
Ham's F12 culture medium, forskolin, isobutylmethylxanthine (IBMX), calphostin C, U73122, 1–10 phenanthrolin, Tri reagent, bovine serum albumin (BSA), trichloroacetic acid (TCA) and trifluoroacetic acid (TFA) were purchased from Sigma (St Louis, MO, USA). Dulbecco's modified Eagle's medium (DMEM), l-glutamine, HEPES, the antibiotic–antimycotic solution, trypsin and agarose were obtained from Gibco Invitrogen Corporation (Grand Island, NY, USA). H89 (N-[2-(P-bromocinnamylamino) ethyl]-5-iso-quinolinesulfonamide) was from Alexis (Laufelfingen, Switzerland). U0126 was from Calbiochem (San Diego, CA, USA). Fetal bovine serum (FBS) was from Dutscher (Brumath, France). Acetonitrile was from Carlo Erba (Val de Reuil, France). Somatostatin-14 was provided by Dr I. Fermé (Sanofi Synthelabo, Le Plessy Robinson, France). Octreotide (SMS 201–995) was provided by Dr R. Maurer (Sandoz, Basel, Switzerland). The selective sst agonists L779-591, L779-976, L797-778 and L803-087 were given by Dr R. M. Freidinger (Merck Research Laboratories, West Point, PA, USA). [Tyr°]-ODN was from Neosystem (Strasbourg, France). Na125I 2000 Ci/mmol was from Amersham International (Les Ulis, France). Rat ODN and rat TTN were synthesized by using the standard Fmoc procedure as previously described (Leprince et al. 1998). All other reagents were of A grade purity.
Secondary cultures of astrocytes were prepared from the brain of Wistar rat as previously described (Brown and Mohn 1999). Briefly, cerebral hemispheres from newborn rats were collected in culture medium consisting of DMEM/Ham's F12 (2/1, v/v) supplemented with 2 mm l-glutamine, 1% insulin, 0.4% glucose, 5 mm HEPES and 1% of the antibiotic–antimycotic solution. The tissues were dissociated mechanically with a syringe equipped with a 1-mm gauge needle, and filtered through a 100-µm sieve (Falcon, Franklin Lakes, NJ, USA). Dissociated cells were resuspended in culture medium supplemented with 10% FBS, plated in 150-cm2 flasks (Techno Plastique Products, Trasandinger, Switzeland) and incubated at 37°C in a 5% CO2/95% O2 atmosphere. When cultures were confluent, astrocytes were isolated by shaking mixed glial cultures overnight with an orbital agitator (KS 15, Bühler, Ratingen, Germany). Adhesive cells were detached by trypsination and preplated for 2 min to discard contaminating oligodendrocytes and microglial cells. Then, the non-adhering cells were harvested and plated on Petri dishes (Dutscher, Brumath, France) at a density of 0.4 × 106 cells/mL. For measurement of endozepine secretion, cells were cultured in 60-mm dishes. For measurement of cAMP and DBI mRNA levels, cells were seeded in 35-mm dishes. After 5 days (DIV5), the proportion of astrocytes was higher than 98% of total cells as determined by staining with antibodies against glial fibrillary acidic protein.
Measurement of DBI mRNA
DIV5 cells were incubated for 24 h at 37°C with fresh serum-free medium in the absence or presence of test substances. At the end of the incubation, the medium was removed, the cells were washed once with phosphate-buffered saline (0.1 m, pH 7.4), and total RNA was extracted by the acid guanidinium thiocyanate–phenol–chloroform method (Chomczynski and Sacchi 1987) using Tri reagent. RNA samples were treated with RNase-free DNase I and approximately 1 µg of total RNA was reverse transcribed by superscript II reverse transcriptase (50 U/µL, Invitrogen, Cergy Pontoise, France) using random hexanucleotides (50 µm) as primers. Real-time RT–PCR was performed on 15 ng of total cDNA using 1 × SYBR green universal PCR Master mix (Applied Biosystems, Foster City, CA, USA) composed of dNTPs, MgCl2 and AmpliTaq Gold DNA polymerase, and 300 nm of the forward (5′-TGCTCCCGCGCTTTCA-3′) and reverse (5′-CTGAGTCTTGAGGCGCTTCAC-3′) DBI primers (Proligo, Paris, France). DBI cDNA was first heated at 50°C for 2 min and 95°C for 10 min, followed by 40 reaction cycles at 95°C for 15 s and 60°C for 1 min, using the ABI Prism 7000 sequence detection system (Applied Biosystems). After PCR, the reaction products were analyzed on a 3% agarose gel to check the size of the amplified DNA. The amount of DBI cDNA in each sample was calculated by the comparative threshold cycle (Ct) method and expressed as 2exp(–Ct) using glyceraldehyde-3-phosphate dehydrogenase (GAPDH) as an internal control.
Measurement of ODN-related peptide release
DIV5 cells were incubated for 5–90 min at 37°C with fresh serum-free medium in the absence or presence of test substances. Culture media were collected and filtered on Sep–Pak C18 cartridges (Waters, Saint-Quentin en Yvelines, France). Bound material was eluted with 50% (v/v) acetonitrile/water containing 0.1% (v/v) TFA and dried by vacuum centrifugation (Speed Vac concentrator, Savant, Hicksville, NY, USA). The dry extracts were kept at 4°C until high-performance liquid chromatography (HPLC) analysis or direct radioimmunoassay (RIA). The cell pellets were used for measurement of protein by the Lowry method.
The dry samples were resuspended in phosphate buffer (0.1 m, pH 8) containing 0.1% Triton X-100 and the concentrations of ODN-like immunoreactivity (ODN-LI) were quantified by RIA using an antiserum raised against synthetic rat ODN (Tonon et al. 1990). The ODN antiserum exhibited 100% cross-reactivity with rat TTN. [Tyr°]-ODN was iodinated using the chloramine-T procedure (Vaudry et al. 1978) and purified on a Sep – Pak C18 cartridge. The final dilution of the ODN antiserum was 1 : 30 000 and the total amount of tracer was 6000 cpm/tube. The antibody-bound ODN fraction was precipitated by adding polyethylene glycol 8000 (20%, w/v) and counted in a gamma counter (LKB Wallac, Evry, France).
The ODN-like peptides contained in incubation media were characterized by HPLC analysis using a Lichrosorb C18 column (250 × 4.6 mm, Touzart et Matignon, Courtaboeuf, France) equilibrated with 20% acetonitrile/water/TFA (20 : 79.9 : 0.1, v/v/v). Dry extracts were diluted in 800 µL 0.1% TFA and eluted at a flow rate of 1 mL/min using the gradient of acetonitrile shown in Fig. 6. One-mL fractions were collected, evaporated and radioimmunoassayed in duplicate. Synthetic rat ODN and TTN (2 µg each), chromatographed in the same conditions, eluted with retention times of 28 and 46 min, respectively.
Measurement of cAMP production
Cultured cells were preincubated for 30 min in serum-free medium containing 10−4m IBMX to inhibit phosphodiesterases and 10−5m forskolin, and then incubated in the same solution, in the absence or presence of test substances. The incubation was stopped by removing the medium and adding 10% (w/v) ice-cold TCA. The cells were homogenized and centrifuged (14 000 g, 4°C, 10 min). The supernatant was collected, washed three times with 1 mL water-saturated diethyl ether, and evaporated overnight. The pellets were used for measurement of protein concentration by the Lowry method. The dry extracts were reconstituted in 1 mL sodium acetate buffer (0.05 m, pH 5.8) and cAMP content was measured by using a commercial RIA kit (RPA 509; Amersham International, Arlington Heights, IL, USA).
All values presented in the figures are means ± SEM. Data were analyzed using anova followed by Bonferroni's test to determine statistical differences between control and experimental values.
Effect of somatostatin on DBI mRNA level
PCR analysis using specific primers for rat DBI and GADPH yielded amplification products with the expected sizes of 94 bp (DBI) and 71 bp (GADPH; Fig. 1a). Quantitative RT–PCR showed that treatment of cultured astrocytes for 24 h with somatostatin (10−6m) provoked a decrease of DBI mRNA concentration without altering the level of GAPDH mRNA (Fig. 1b). Somatostatin induced a concentration-dependent diminution of DBI mRNA levels with an IC50 value of 10−10m and a maximum effect at a concentration of 10−8m(Fig. 2a). The selective sst4 agonist L803-087 also induced a concentration-dependent reduction of DBI mRNA levels with an IC50 value of 0.6 × 10−10m (Fig. 2b), whereas the selective sst1, sst2 and sst3 receptor agonists L779-591, L779-976 and L797-778 had no effect on DBI mRNA level even at high concentrations (Fig. 2b). The selective MEK inhibitor U0126 reduced by 38% DBI mRNA level in rat astrocytes. In addition, somatostatin was unable to further reduce DBI mRNA level in the presence of U0126 (Fig. 3).
Effect of somatostatin on endozepine release
Incubation of cultured astrocytes with graded concentrations of somatostatin (10−12 to 10−6m) for 30 min induced a concentration-dependent inhibition of the release of ODN-LI (Fig. 4a). The inhibitory effect of somatostatin was already significant at a dose of 10−11m and the IC50 value was 0.6 × 10−10m. L779-591, L779-976 and L803-087 also induced a concentration-dependent inhibition of the release of ODN-LI, with IC50 values of about 0.08 × 10−10m (Fig. 4b) and maximal inhibition was observed at concentrations of 0.1–0.3 × 10−10m.
Time-course experiments showed that somatostatin (10−9m) significantly reduced the secretion of ODN-LI within 20 min and reached a maximum (− 50%) after 30 min of incubation (Fig. 5a). Thereafter, the action of somatostatin gradually declined and vanished 1 h after the onset of peptide administration. Addition of a novel dose of somatostatin (10−9m) after 30, 50 or 60 min of incubation restored the inhibitory effect on ODN release (Fig. 5a). In contrast, octreotide (10−9m) induced a prolonged inhibitory effect that lasted for more than 70 min. Similarly, phenanthrolin (10−4m), a specific inhibitor of endopeptidases 24.15 and 24.16, maintained at least for 60 min the inhibitory effect of somatostatin on the release of ODN-LI (Fig. 5b).
Reversed-phase HPLC analysis combined with RIA quantification of conditioned media from cultured astrocytes, revealed the existence of a major immunoreactive form which coeluted with synthetic rat TTN (46 min; Fig. 6a). Two additional minor forms were resolved, one co-eluting with synthetic rat ODN (26 min) and a more hydrophobic compound exhibiting a retention time of 73 min. Incubation of cultured rat astrocytes with somatostatin (10−9m, 30 min) reduced by approximately 45% the amount of all immunoreactive peptides (Fig. 6b).
Involvement of the PKA pathway in the effect of somatostatin on endozepine release
Exposure of cultured astrocytes to graded concentrations of somatostatin (10−12 to 10−6m, 30 min) induced a concentration-dependent inhibition of forskolin-induced cAMP formation with an IC50 value of 0.06 × 10−10m. The maximal effect (− 40%) was observed at a dose of 10−10m(Fig. 7a). The selective sst1, sst2 and sst4 agonists L779-591, L779-976 and L803-087 produced similar dose–response curves with IC50 values of 0.09 × 10−10m, 0.05 × 10−10m and 0.7 × 10−10m, respectively (Fig. 7b). Time-course experiments showed that somatostatin (10−9m) caused a decrease in cAMP content in the cells that reached a maximum after 15 min of treatment. After a 1-h exposure to somatostatin, the intracellular level of cAMP returned to the control value (Fig. 7c).
Somatostatin (10−9m) significantly reduced forskolin-evoked stimulation of the secretion of ODN-LI (Fig. 8). The selective protein kinase A (PKA) inhibitor H89 (2 × 10−5m) mimicked the inhibitory effect of somatostatin on the secretion of ODN-LI, and the actions of somatostatin and H89 were not additive (Fig. 8). In contrast, the PLC inhibitor U73122 (10−5m), the protein kinase C (PKC) activator PMA (10−6m), or the PKC inhibitor calphostin C (10−7m) did not affect basal nor somatostatin-evoked inhibition of ODN-LI secretion (Fig. 9).
It has previously been shown that endozepines are synthesized by glial cells (Lamacz et al. 1996) and that the release of DBI-related peptides by cultured rat astrocytes is a regulated process (Patte et al. 1999; Masmoudi et al. 2003). It has also been found that astroglial cells possess a high density of somatostatin-binding sites (Mentlein et al. 1990; Krisch 1994; Hosli and Hosli 2000). In the present study, we demonstrate that somatostatin, acting through the mitogen-activated protein kinase (MAPK) and/or the adenylyl cyclase/PKA transduction pathways, reduces the levels of DBI mRNA and inhibits the release of endozepines in cultured rat astrocytes.
Although it has been reported that the DBI gene exhibits the hallmarks of a typical housekeeping gene (Mandrup et al. 1992; Kolmer et al. 1993; Swinnen et al. 1996b), several studies have shown that the DBI promoter encompasses consensus sequences for regulatory elements such as sterol regulatory element (SRE), nuclear factor-Y (NF-Y) and glucocorticoid-response element (GRE) (Mandrup et al. 1992; Swinnen et al. 1998). It has also been found that insulin and androgens stimulate DBI gene expression in the 3T3-L1 pre-adipocytes (Hansen et al. 1991) and in the LNCaP prostatic (Swinnen et al. 1996a) cells lines, respectively. Here, we show that subnanomolar concentrations of somatostatin markedly reduced DBI mRNA levels in rat astrocytes, indicating that, in the CNS as in peripheral tissues, the expression of the DBI gene is finely regulated.
Rat cortical astrocytes express different subtypes of receptors for somatostatin, i.e. sst1, sst2 and sst4 and occasionally sst3 receptor (Feindt et al. 1995). The present study shows that the selective sst4 receptor agonist L803-087 dose-dependently reduced the concentration of DBI mRNA while the sst1, sst2 and sst3 agonists L779-591, L779-976 and L797-778 were inactive. It thus appears that the down-regulatory effect of somatostatin on DBI mRNA level in cultured rat astrocytes is most likely mediated through the sst4 receptor subtype. While sst1, sst2, sst3 and sst4 receptors share common signaling pathways including inhibition of adenylyl cyclase and activation of tyrosine phosphatase (Reisine and Bell 1995; Meyerhof 1998; Csaba and Dournaud 2001), the sst4 receptor has been shown to act also on the MAPK/MEK transduction pathway (Sakanaka et al. 1994). Because the MEK inhibitor U0126 reduced by itself DBI mRNA levels and because the effects of somatostatin and U0126 were not additive, it appears that the action of somatostatin on DBI mRNA level is mediated through the ERK/MAPK signalling cascade.
We have then found that subnanomolar concentrations of somatostatin induced a dose-dependent inhibition of ODN-LI release by cultured astrocytes. The selective sst1, sst2 and sst4 agonists all mimicked the inhibitory effect of somatostatin on endozepine secretion. The fact that all three compounds were ≈10 times more potent than somatostatin itself is consistent with binding data that indicate that L779-976 and L803-087 exhibit higher affinities for sst2 and sst4 receptors, respectively, than the native peptide (Blake 2001; Pfeiffer et al. 2001; Cervia et al. 2003). The present study thus reveals that, in glial cells, sst1, sst2 and sst4 receptors act redundantly to mediate somatostatin inhibition of endozepine release. The expression of multiple sst isoforms in a single cell population is not unusual and it is indeed rather common to find targets cells equipped with more than one sst receptor subtype (Patel et al. 1994; Khare et al. 1999). Because the effect of somatostatin on DBI mRNA level is largely mediated through the sst4 receptor subtype, these data suggest that somatostatin can either simultaneously inhibit DBI gene expression and endozepine secretion (through activation of sst4) or selectively reduce endozepine release (through activation of sst1 and sst2). Besides, it has been recently shown that the sst1 and sst5 receptors can form heterodimers when the two receptor subtypes are expressed in CHO cells, and that dimerization alters the functional properties of the receptor including ligand binding affinity and agonist-induced receptor up-regulation and internalization (Rocheville et al. 2000; Patel et al. 2002). Whether sst1, sst2 and sst4 receptors may form heterodimers in rat glial cells remains to be determined.
Previous studies have shown that agonist stimulation induces phosphorylation and internalization of various somatostatin receptors (Roosterman et al. 1997; Roth et al. 1997; Boudin et al. 2000; Stroh et al. 2000; Liu and Schonbrunn 2001). In particular, studies conducted on cultured glial cells have revealed internalization of a fluorescent analog of somatostatin within 30 min (Stroh et al. 2000). Thus, the transient effect of somatostatin on endozepine secretion that we have initially observed during the time-course experiments was first attributed to receptor internalization. However, subsequent studies clearly established that the attenuation of the inhibitory effect of somatostatin on endozepine release during prolonged administration of the neuropeptide could be accounted for by degradation of the ligand rather than by down-regulation of the receptors: (i) additional application of somatostatin 30 min after the onset of the peptide administration restored the inhibitory effect of somatostatin; (ii) the analogue octreotide induced a sustained inhibition of endozepine release; and (iii) the metalloprotease inhibitor phenanthrolin maintained the inhibitory activity of somatostatin on endozepine release for at least 60 min. Consistent with these observations, it has been previously reported that somatostatin is rapidly inactivated by endopeptidases 24.15 and 24.16, two metalloproteases that are sensitive to phenanthrolin (Mentlein and Dahms 1994). In particular, it has been shown that conditioned media from cultured astrocytes cleave somatostatin at the Phe6–Phe7 and Thr10–Phe11 bonds (Veber et al. 1979; Mentlein and Dahms 1994). Altogether, these data indicate that the decrease of the response of astrocytes to somatostatin can be ascribed to the inactivation of the neuropeptide by endopeptidases released from cultured cells.
We have previously shown that cultured glial cells release at least two forms of endozepines, i.e. the triakontatetraneuropeptide TTN and the octadecaneuropeptide ODN (Patte et al. 1999; Masmoudi et al. 2003). TTN acts preferentially on peripheral-type benzodiazepine receptors (Slobodyansky et al. 1989) while ODN is a selective ligand for central-type benzodiazepine receptors (Ferrero et al. 1986) so that the two peptides exert distinct activities on the CNS. The present study reveals that somatostatin inhibited similarly the release of TTN and ODN, thus indicating that it had no effect on the processing of DBI in rat astrocytes.
It has been demonstrated that somatostatin inhibits adenylyl cyclase activity in rat astrocytes (Evans et al. 1984; Grimaldi et al. 1997) and human glioma cells (Feindt et al. 1995; Lamszus et al. 1997). Here, we show that the concentration-dependent and time-course effects of somatostatin on cAMP formation were strictly parallel to those observed on endozepine release. Two additional observations indicate that the inhibitory effect of somatostatin on endozepine secretion can be accounted for by inhibition of the adenylyl cyclase/PKA signaling pathway. First, somatostatin completely blocked forskolin-evoked endozepine release. Second, the PKA inhibitor H89 mimicked the inhibitory effect of somatostatin, and the actions of H89 and somatostatin on ODN-LI secretion were not additive. We have recently shown that the neuropeptide PACAP stimulates endozepine release from rat astrocytes through activation of the adenylyl cyclase/PKA transduction pathway (Masmoudi et al. 2003). It has also been reported that the inhibitory effect of somatostatin on IL6 release from cultured glial cells is mediated through inhibition of cAMP formation (Grimaldi et al. 1997). Taken together, these data reveal the crucial role of the adenylyl cyclase/PKA pathway in the control of the secretory activity of rat astrocytes.
In conclusion, the present study has demonstrated that somatostatin reduced DBI mRNA concentration and inhibits endozepine release by cultured rat astrocytes. The effect of somatostatin on DBI mRNA level can be ascribed to activation of sst4 receptors negatively coupled to the MAPK signalling cascade while the effect of somatostatin on endozepine release is mediated through activation of sst1, sst2 and sst4 receptors negatively coupled to the adenylyl cyclase/PKA pathway. This study has also shown that rat astroglial cells produce endoproteases that inactivate somatostatin and thus attenuate its inhibitory action on endozepine secretion.
The authors wish to thank Mrs Hugette Lemonnier and Mr Gérard Cauchois for skillful technical assistance. This work was supported by INSERM (U413) and the Conseil Régional de Haute-Normandie. OM was the recipient of a fellowship from the French and Tunisian Ministries of Research. TT was the recipient of a fellowship from the Chinese Ministry of Research. AR was the recipient of a fellowship from the French Ministry of Research.