Address correspondence and reprint requests to Hadassah Tamir, Department of Neuroscience, New York State Psychiatric Institute, 1051 Riverside Drive, New York, NY 10032, USA. E-mail: firstname.lastname@example.org
Positron emission tomography studies in major depression show reduced serotonin (5-HT)1A receptor antagonist-binding potentials in many brain regions including occipital cortex. The functional meaning of this observation in terms of signal transduction is unknown. We used postmortem brain samples from depressed suicide victims to examine the downstream effectors of 5-HT1A receptor activation. The diagnosis was established by means of psychological autopsy using Diagnostic and Statistical Manual of Mental Disorders (DSM) III-R criteria. Measurements of [35S]GTPγS binding to Gαi/o in the occipital cortex of suicide victims and matched controls revealed a blunted response in suicide subjects and a decrease in the coupling of 5-HT1A receptor to adenylyl cyclase. No significant group differences were detected in the expression levels of Gαi/o, Gαq/11 or Gαs proteins, or in the activity of cAMP-dependent protein kinase A. Studies of a parallel transduction pathway downstream from 5-HT1A receptor activation demonstrated a decrease in the activity of phosphatidylinositol 3-kinase and its downstream effector Akt, as well as an increase in PTEN (phosphatase and tensin homolog deleted on chromosome 10), the phosphatase that hydrolyzes phosphatidylinositol 3,4,5-triphosphate. Finally, the activation of extracellular signal-regulated kinases 1 and 2 was attenuated in suicide victims. These data suggest that the alterations in agonist-stimulated 5-HT1A receptor activation in depressed suicide victims are also manifest downstream from the associated G protein, affecting the activity of second messengers in two 5-HT1A receptor transduction pathways that may have implications for cell survival.
Although several findings indicate that levels of the neurotransmitter serotonin (5-HT) are reduced in depressed suicide victims (Mann 1998), the physiological consequences of such a reduction for 5-HT-receptor-activated downstream signaling are less well described. Autoradiographic studies on brains obtained from suicide victims revealed an increased density of postsynaptic 5-HT1A receptor agonist binding in the ventrolateral prefrontal cortex (PFC) (Yates and Ferrier 1990; Arango et al. 1995; Lowther et al. 1997). In contrast, positron emission tomography studies of depressed subjects revealed lower 5-HT1A antagonist-binding potentials in several brain regions, including frontal, temporal and occipital cortex (Drevets et al. 1999; Sargent et al. 2000). Radioligand binding studies, however, can only measure available ligand-binding sites, and do not provide information about signal transduction pathways that exponentially amplify the signals of ligand-induced receptor activation. Thus, measurements of the extent to which 5-HT1A receptors are responsive to agonist stimulation to activate second messengers may provide a more meaningful index of serotonergic system function.
The 5-HT1A receptor is a member of the 5-HT1 family which is composed of G protein-coupled receptor isoforms that inhibit the activity of adenylyl cyclase (De Vivo and Maayani 1987; Hoyer et al. 1994). Activation of 5-HT1A receptors decreases Ca2+ conductance (Penington et al. 1991), increases K+ conductance (Andrade et al. 1986), and activates mitogen-activated protein (MAP) kinases (extracellular signal-regulated kinases; ERKs) (Cowen et al. 1996) via pertussis toxin (PTX)-sensitive G proteins (Gαi/Gαo). In addition, 5-HT1A receptors can couple to Gαz, a PTX-insensitive G protein, to increase the secretion of some endocrine hormones (Serres et al. 2000). A novel regulation of this receptor by phospholipid-derived signaling components has been described (Evans et al. 2001).
Little is known about potential changes in second messenger activities associated with the 5-HT1A receptor activation in brains of patients with major depressive disorder (MDD) and suicide victims. Several studies have begun to functionally assess serotonergic neurotransmission by measuring signal transduction molecules in postmortem tissues. For example, Cowburn et al. (1994) studied postmortem frontal cortex of suicide victims and showed that G protein as well as the forskolin-induced adenylyl cyclase activity changes were blunted compared with matched controls despite unaltered expression levels of Gαs and Gαi protein. Reiach et al. (1999) confirmed these observations by demonstrating reduced adenylyl cyclase immunolabeling and activity in the temporal cortex of depressed suicide victims. More recent studies, however, which measured agonist-stimulated [35S]GTPγS binding in the PFC of suicide victims with MDD, failed to demonstrate any difference between suicide victims and matched controls in agonist-stimulated 5-HT1A receptor-mediated G protein activation and levels of G protein expression (Gonzalez-Maeso et al. 2002). In contrast, Pacheco et al. (1996) demonstrated higher Gαs levels in the PFC of depressed suicide victims. Finally, a study that examined subjects with mood disorders and suicide victims with no history of medication before death found no differences in temporal and occipital cortical Gαs levels, and only a trend towards blunted forskolin-stimulated adenylyl cyclase activity (Dowlatshahi et al. 1999).
The main downstream effector of adenylyl cyclase is cyclic AMP (cAMP)-dependent protein kinase A (PKA) activity, which has been studied in suicide victims (Dwivedi et al. 2002). [3H]cAMP binding to PKA estimates the level of the regulatory subunit of the kinase and assay of kinase activity measures the level of the catalytic subunit of the enzyme. Suicide victims with a history of MDD express lower levels of both regulatory and catalytic subunits of PKA in the PFC compared with controls. In bipolar disorder, however, Rahman et al. (1997) demonstrated lower [3H]cAMP binding whereas Fields et al. (1999) demonstrated greater PKA activity in the PFC.
Receptor-stimulated transduction pathways can also be activated by the Gβγ subunits. Such a pathway was first discovered for adrenergic receptors (Touhara et al. 1995) and includes activation of phosphatidylinositol 3-kinase (PI3-K) and its downstream effector, protein kinase B (Akt). PI3-K phosphorylates the D3 position of the inositol ring of phosphoinositides, generating intracellular lipid second messengers. This lipid kinase is involved in numerous physiological functions (Vanhaesebroeck et al. 1997; Kandel and Hay 1999; Scheid and Woodgett 2001). Regulation of this pathway is complex and includes positive (Fillipa et al. 1999) and negative (Kim et al. 2001; Lou et al. 2002) regulation by cAMP. Another important negative regulation of Akt is achieved by phosphatase and tensin homolog deleted on chromosome 10 (PTEN), which dephosphorylates phosphatidylinositol 3,4,5-triphosphate (PIP3), the product of PI3-K stimulation, to form phosphatidylinositol 4,5-biphosphate, and thus opposes the action of PI3-K (Choi et al. 2002; for review see Cantley and Neel 1999) and thereby decreases Akt activity. The PI3-K/Akt pathway has been implicated in numerous physiological functions including activation of the ERKs via the 5-HT1A receptor (Cowen et al. 1996). Moreover, PI3-K/Akt is required for cell survival (Franke et al. 1997; Fruman et al. 1998). The significance of the dual signaling pathways via the 5-HT1A receptor is not clear and the mechanisms involved in coordinating the two signal pathways remain to be elucidated.
In the present study, we examined distinct activities of second messengers activated by 5-HT1A receptors. We used occipital cortical tissues obtained from suicide victims and matched controls to measure: (1) 5-HT1A receptor-activated coupling to G protein; (2) specific G proteins; (3) sensitivity to PTX of receptor-activated G protein; (4) forskolin- and receptor-modulated activity of adenylyl cyclase; (5) total activity of PKA (in the presence of cAMP); and (6) whether other signal transduction pathways activated by 5-HT1A receptors, namely PI3-K, Akt, PTEN and MAP kinase, are altered in suicide victims.
Materials and methods
Unless otherwise stated the reagents used were purchased from Sigma-Aldrich (St Louis, MO, USA). Forskolin was purchased from Calbiochem (La Jolla, CA, USA). PTX was purchased from List Biological Laboratories (Campbell, CA, USA). [Adenylate-32P]NAD (30 Ci/mmol), [γ32P]ATP (3000 Ci/mmol) and [35S]GTPγS (1250 Ci/mmol) were purchased from Perkin Elmer Life Sciences (Boston, MA, USA). Phosphatidylinositol was purchased from Avanti Polar Lipids, Inc. (Alabaster, AL, USA).
Subjects, tissue collection and dissection
Tissue was collected in accordance with protocols approved by the relevant institutional review boards. All subjects died suddenly and had a postmortem interval (time from death to freezing) of less than 26 h. The next of kin consented to tissue collection, review of relevant records and a psychological autopsy. At least one informant per case, for both suicide victims and controls, agreed to be interviewed for the purpose of a psychological autopsy, which generated an axis I and axis II Diagnostic and Statistical Manual of Mental Disorders (DSM) III-R diagnosis (Kelly and Mann 1996). Further details about the psychological autopsy procedure have been described previously (Mann et al. 2000). Toxicological screens, performed on all subjects, ruled out recent consumption of psychotropic medication.
The occipital cortex (containing Brodmann areas 17, 18 and 19) was dissected frozen by a neuroanatomist, and the cortical gray matter was isolated and stored at − 80°C until membrane preparation. Samples from six controls known to be free from psychiatric illness, without a family history of an axis I psychiatric disorder and dying from a cause other than suicide (age range 24–66 years), were compared with those from a matched group of six subjects who had a lifetime diagnosis of a major depressive episode (including one subject with bipolar disorder) and died by suicide (age range 19–71 years). The groups were matched for age, sex and postmortem intervals (Table 1). Linear regression analysis was carried out to calculate the correlation of biochemical data with postmortem interval, gender and sex. No correlations were observed ( p > 0.15). All cases were free of significant neuropathology. The samples were coded to mask the investigators to the diagnostic group.
Table 1. Demographics, diagnoses and toxicology of subjects in study
Preparation of membrane fractions from human occipital cortex
This procedure is a modification of a previously published method (Adlersberg et al. 2000). The intact cortical gray tissue was kept frozen and homogenized just before the assay. Samples obtained from controls and suicide victims were assayed simultaneously and in triplicate. Variability within samples was around 10–15%. Tissue was homogenized in 20 volumes of homogenization buffer containing 50 mm Tris-HCl (pH 7.4), 3 mm MgCl2, 1 mm EGTA, 1 mm dithiothreitol (DTT), and a cocktail of protease inhibitors at 1 : 2000 dilution. The crude homogenate was centrifuged at 180 g for 5 min and the supernatant was incubated for 5 min at 30°C to remove endogenous 5-HT, and then centrifuged at 100 000 g for 10 min at 4°C. The pellet was suspended in the same buffer and recentrifuged. The final pellet was suspended in assay buffer containing 50 mm Tris-HCl (pH 7.4), 5 mm MgCl2, 0.2 mm EGTA, 100 mm NaCl, 1 mm DTT and a cocktail of protease inhibitors. For the study of [35S]GTPγS binding, membranes were stored at − 70°C for no longer than 5 days before the assay. The assay of adenylyl cyclase was carried out on the same day.
Agonist-stimulated [35S]GTPγS binding to membranes
These experiments were carried out as described previously (Adlersberg et al. 2000). Briefly, membrane fractions (10–20 µg) were preincubated for 5 min at 30°C with the indicated concentration of 5-HT1A agonist (+/–)-8-hydroxy-dipropylaminotetralin (8-OH-DPAT) in the presence of 30 µm GDP in a final volume of 0.5 mL assay buffer (see above). [35S]GTPγS (0.05 nm; 1250 Ci/mmol) was added, and the incubation was continued for 45 min. Incubations were terminated by rapid filtration through glass filters (GF/B), and bound radioactivity was determined by liquid scintillation spectrometry. Basal activity was determined in the absence of agonist, and non-specific binding was estimated in the presence of unlabeled GTPγS. The specificity of the agonist was tested by incubating the membranes with a 5-HT1A-specific antagonist N-[2-[4-(2-methoxyphenyl)-1-piperazinyl]ethyl]-N-2-pyridinylcyclohexanecarboxyamide (WAY-100635) (0.1–50 nm) for 15 min at 30°C before addition of 8-OH-DPAT (100 nm).
Quantitation of G proteins by western blot analysis
Brain tissues (0.2 g) from controls and suicide victims were homogenized in a buffer containing 50 mm Tris HCl (pH 7.4), 3 mm MgCl2, 1 mm EGTA, 1 mm DTT and a cocktail of protease inhibitors (1 : 2000). The homogenate was centrifuged at 100 000 g for 10 min at 4°C. The pellet was washed and resuspended in the same buffer. Proteins (20 µg) were separated on 10% sodium dodecyl sulfate (SDS)–polyacrylamide gels and electroblotted on to nitrocellulose membranes. The proteins were visualized by staining for 2 min in Ponceau S solution (0.02% in 0.3% trichloroacetic acid) and membranes were incubated with primary antibodies [rabbit polyclonal anti-GαsL+S, anti-Gαi1 and Gαi2 which did not cross-react with Gαo, anti-Gαq/11 (1 : 500; Santa Cruz Biotechnology; Santa Cruz, CA, USA)] and anti-β-actin monoclonal antibody (1 : 5000) at 4°C overnight. Reactive bands were visualized with a 1 : 2500 dilution of horseradish peroxidase (HRP)-labeled goat anti-rabbit and anti-mouse secondary antibodies (Santa Cruz Biotechnology) in conjunction with enhanced chemiluminescence (ECL; Pierce, Rockfort, IL, USA). In all western blot analyses actin was added as a standard, and the density of the specific protein tested was presented as a ratio of its optical density to that of actin. This procedure corrects for loading variability. Variability within samples was around 10%.
Quantitation of Gαi/o protein by PTX-catalyzed adenosyl diphosphate ribosylation
These experiments were carried out as described previously (Carty 1994). Briefly, activation of PTX was achieved by incubating the toxin with 50 mm Tris-HCL (pH 7.4), 20 mm DTT, 0.1% SDS and 0.1% bovine serum albumin (BSA) at 30°C for 30 min. Activated PTX (1 ng/µL) was added to membranes (10 µg) in 50 µL of buffer containing 25 mm Tris-HCl (pH 7.4), 1 mm EDTA, 10 mm DTT, 0.005% SDS, 10 mm thymidine, 0.5 mm benzamidine, 10 µg/µL trypsin inhibitor, protease inhibitors (1 : 1000), 2.5 µm NAD and 2 µCi [adenylate-32P]NAD (30 Ci/mmol). The mixture was incubated for 30 min at 30°C. The reaction was stopped by the addition of SDS sample buffer and boiling for 5 min. The reaction products were separated on 11% SDS–polyacrylamide gels. Gels were stained with Coomassie blue R-250 (Bio-Rad, Hercules, CA, USA) and dried. Specific phosphorylated bands were identified by their molecular weight using autoradiography.
Assay of adenylyl cyclase activity
Crude membranes (5 µg protein) of controls and suicide victims were suspended in a 200-µL incubation mixture containing 50 mm Tris-HCl (pH 7.4), 5 mm MgCl2, 0.1% BSA, 0.1 mm EDTA, 1 mm EGTA, 1 mm DTT, 0.1 mm isobutylmethylxanthine (IBMX; inhibitor of cAMP phosphodiesterase), 5 mm phosphocreatine, 50 U/mL creatine phosphokinase, 10 µm GTP and a cocktail of protease inhibitors (1: 2000). Adenylyl cyclase was activated by the addition of forskolin (1–50 µm) in the presence or absence of 5-HT1A receptor stimulation by the agonist 8-OH-DPAT (1–100 µm) to measure the degree of adenylyl cyclase inhibition. Basal activity was measured in the absence of agonist or forskolin. The reaction was initiated by the addition of 0.5 mm ATP, continued for 15 min at 30°C, and stopped by the addition of 20 µL 1 N HCl (final concentration 0.1 N). Samples were centrifuged at 14 000 g for 10 min at 4°C. Supernatants were diluted 1 : 40 and assayed for cAMP levels, in duplicate, using a commercial enzyme immunoassay kit (Assay Designs, Ann Arbor, MI, USA).
Assay of PKA activity
Tissues from controls and suicide victims were homogenized in 10 volumes of lysis buffer [20 mm 2-(N-morpholino) ethanesulfonic acid (pH 6.5), 1 mm EDTA, 50 mm NaCl, 0.2% Triton X-100 and 1 mm IBMX]. The homogenates were centrifuged at 14 000 g for 10 min and the recovered cytosolic fractions were assayed for PKA activity by measuring the phosphorylation of an exogenous substrate (a synthetic peptide named kemptide) by [γ32P]ATP in the presence of increasing concentrations of cAMP (total activity) or in the absence of cAMP (basal activity). The assay buffer contained 20 mm 2-(N-morpholino) ethanesulfonic acid (pH 6.5), 0.4 mm EDTA, 50 mm NaCl, 2.5 mm Mg acetate, 0.2 mm kemptide, 0.15% BSA, 0.25 µCi [γ32P]ATP (3000 Ci/mmol) and protease inhibitors. The PKA assay (100 µL) was initiated by the addition of 20 µg cytosolic proteins and continued for 5 min at 30°C. The reaction was stopped by spotting aliquots (20 µL) on to squares of phosphocellulose paper. The paper was washed three times with cold 0.5% H3PO4, once with acetone, air-dried, and counted in a scintillation counter to measure 32P incorporation. PKA activity was expressed as nmoles 32P transferred to kemptide per minute per milligram protein.
Immunolabeling of Akt, PTEN and p44/42 MAP kinase (ERK1/2)
Tissues were homogenized in lysis buffer containing 20 mm Tris-HCl (pH 7.4), 150 mm NaCl, 1 mm EGTA, 1 mm EDTA, 1% Triton X-100, phosphatase inhibitors (2.5 mm sodium pyrophosphate, 1 mmβ-glycerophosphate, 1 mm sodium vanadate) and a cocktail of protease inhibitors (1: 2000). The lysate was centrifuged at 14 000 g for 30 min at 4°C. Aliquots of the lysate (25 µg protein) were separated on 10% SDS–polyacrylamide gels and electroblotted on to nitrocellulose membranes. The proteins were then visualized by staining for 2 min in Ponceau S solution (0.02% in 0.3% trichloroacetic acid), and membranes were incubated with rabbit polyclonal anti-p44/p42 MAP kinase, anti-PTEN, anti-Akt (1 : 1000; Cell Signaling Technology, Beverly, MA, USA) and anti-β-actin monoclonal antibody (1 : 5000) at 4°C overnight. Reactive bands were visualized with a 1 : 2500 dilution of HRP-labeled goat anti-rabbit or anti-mouse secondary antibody.
Measurement of activated p44/42 MAP kinase
Lysates for the two experimental groups were prepared as described above. P44/42 MAP kinase activity was determined by western blot analysis using a phospho-specific antibody that detects endogenous levels of p44/42 MAP kinase (ERK1 and ERK2) only when catalytically activated by phosphorylation at Thr202 and Tyr204. Aliquots of lysates (50 µg protein) were separated on 10% SDS– polyacrylamide gels and electroblotted on to nitrocellulose membranes. The membranes were washed and phosphorylated forms of p44/42 MAP kinase were detected by incubating the membrane with a phospho-p44 MAP kinase polyclonal antibody (1 : 1000; Cell Signaling Technology) and anti-β-actin monoclonal antibody overnight at 4°C. Reactive bands were visualized with a 1 : 2500 dilution of HRP-labeled goat anti-rabbit or anti-mouse secondary antibody.
Assay of PI3-K activity
PI3-K activity was measured by a previously published method with minor modification (Liu et al. 2003). Aliquots of lysate (750 µg protein) were immunoprecipitated with a polyclonal antibody directed against the p85 subunit of PI3-K (5 µg/sample; Santa Cruz Biotechnology) and allowed to incubate on ice for 60 min. Protein A/G–agarose (30 µL) was then added, and incubation was continued for 60 min at 4°C. After centrifugation of the agarose–antigen–antibody complex, the pellet was washed with 500 µL lysis buffer [25 mm Hepes (pH 7.5), 150 mm NaCl, 1 mm EDTA, 1 mm EGTA, 1% Triton X-100, 1 mm sodium vanadate and a cocktail of protease inhibitors] followed by washing with 500 µL kinase buffer [30 mm Hepes (pH 7.5) and 30 mm MgCl2]. The complex was suspended in sonicated soybean phosphatidylinositol (20 µg) and 35 µL kinase assay buffer as described above. The reaction was initiated by the addition of 50 µm ATP and 5 µCi [γ32P]ATP (3000 Ci/mmol) in 5 µL kinase buffer, incubated for 10 min at 30°C, and stopped by the addition of 100 µL 1 N HCl. Lipids were extracted with 200 µL chloroform : methanol (1 : 1), spotted on to silica gel G-60 TLC plates, and developed in a mobile phase consisting of chloroform, acetone, methanol, acetic acid and water (40 : 15 : 13 : 12 : 7; v/v/v/v/v). Spots corresponding to phosphatidylinositol 3-phosphate [PI(3)-P] were detected by autoradiography and identified on the basis of their co-migration with a known standard. Kinase activity was quantified by cutting the spots corresponding to PI3-P from the plate and analyzing them by liquid scintillation counting.
Assay of Akt activity
Aliquots of lysates (500 µg protein) were immunoprecipitated by adding 20 µL resuspended immobilized Akt 1G1 monoclonal antibody slurry (Cell Signaling Technology) and incubated with gentle rocking overnight at 4°C. Lysates were then microfuged for 30 s at 4°C and the pellets were washed twice with 500 µL lysis buffer [25 mm Tris (pH 7.5), 150 mm NaCl, 1 mm EDTA, 1 mm EGTA, 1% Triton X-100, 2.5 mm sodium pyrophosphate, 1 mmβ-glycerophosphate, 1 mm sodium vanadate, 100 nm okadaic acid, 100 nm calyculin and a cocktail of protease inhibitors] followed by one wash with 500 µL kinase buffer [25 mm Tris (pH 7.5), 5 mmβ-glycerophosphate, 2 mm DTT, 0.1 mm sodium vanadate and 10 mm MgCl2]. The resulting pellet was resuspended in 40 µL kinase buffer supplemented with 200 µm ATP and 1 µg glycogen synthase kinase-3 (GSK-3) fusion protein (a synthetic substrate), and incubated for 30 min at 30°C. The reaction was terminated with 40 µL 2 × SDS sample buffer, and proteins were separated on 12% SDS–polyacrylamide gels and electroblotted on to nitrocellulose membranes. The membranes were washed and phosphorylated GSK-3α/β was detected by incubating the membrane overnight at 4°C with a phospho-GSK-3α/β(Ser21/9) polyclonal antibody (1 : 1000 dilution). The same membranes were also incubated with anti-Akt (1 : 1000) to ensure that equal amounts of Akt were immunoprecipitated. Reactive bands were visualized with a 1 : 2500 dilution of HRP-labeled goat anti-rabbit secondary antibody. Bound antibodies were visualized on blots by enhanced chemiluminescence.
Data analysis was performed using non-linear regression (Kaleida-Graph; Synergy Software, Reading, PA, USA). All experiments were repeated independently at least three times. In separate analysis, where appropriate, activity differences in individual parameters were analyzed using repeated measures two-way anova followed by a post hoc test (Univariate F-test, Bonferroni post hoc test; adjusted level of significance p < 0.01). For sample analysis, both paired and non-paired tests were performed, and both were found to be significant (non-significant p > 0.05; unless otherwise stated). The relationship between [35S]GTPγS binding, adenylyl cyclase, Akt, PI3-K, PTEN, P44/42MAPK and postmortem interval, gender and sex were determined by linear regression analysis; none was found to be significant. The level of non-significance was chosen as > 0.1). The correlation, if any, between exposure to drugs and the various parameters mentioned above could not be evaluated because no brain tissue was exposed to the same drug. Quantification of enhanced chemiluminescence signals on immunoblots was performed using the NIH image analysis software (version 1.62; NIH, MD, USA). All results are presented as mean ± SEM and when indicated p values were determined by means of an unpaired Student's t-test.
The coupling of 5-HT1A receptor to Gαi/o protein is diminished in depressed suicide victims
We examined 5-HT1A receptor activation of G proteins by measuring agonist-induced increases in [35S]GTPγS binding. Membrane fractions obtained from controls were stimulated with 8-OH-DPAT and the radioactivity incorporated into proteins was measured. Agonist-stimulated binding was dose dependent with an EC50 of 7.5 nm(Fig. 1a). Binding specificity was demonstrated by competition with the specific antagonist WAY-100635, with an IC50 of 12 nm (Fig. 1b). The agonist-stimulated [35S]GTPγS binding was abolished by pretreatment of the membranes with PTX (1 µg/mL) (data not shown), indicating that 5-HT1A receptors are coupled to Gαi/o, and not to Gαz. Binding of [35S]GTPγS was 44 ± 5% lower in suicide victims (Fig. 1c). This difference in [35S]GTPγS binding (p < 0.01) might be explained by a smaller numbers of receptors, a reduction in G protein levels, or a decrease in the coupling of the receptor to G proteins.
Immunolabeling of G protein subunits in the occipital cortex of controls and suicide victims
In order to test whether the levels of Gα proteins in the occipital cortex of suicide victims were altered, we separated and identified the major forms of these proteins present in the brains of suicide victims and controls. Samples containing 20 µg protein were loaded on to the gel to ensure that the chemiluminescent signals were within the linear range. Representative examples of quantification of signals on autoradiograms corresponding to Gαs (45 kDa), Gαi2 and Gαq/11 subunits in the occipital cortex of six paired controls and suicide victims are given in Fig. 2a. Variance within samples was < 10%. No group differences were observed in the immunoreactive levels of all three forms of Gα protein. Low levels of Gαi1 were also detected, but no significant group differences were found (data not shown).
Quantification of Gαi/o protein by PTX-catalyzed adenosyl diphosphate ribosylation
To verify the above results, we measured the levels of Gαi/o by an independent and complementary technique, PTX-catalyzed ADP ribosylation. We took advantage of the specificity of PTX to ADP ribosylate Gαi/o in the presence of [adenylate-32P]NAD to radiolabel the protein. [32P]ADP-ribosylated Gαi/o was separated from other components of the reaction mixture, autoradiographed and quantified. No differences were found in the radiolabeling of Gαi/o in samples from suicide victims compared with controls (Fig. 2b). These results, in conjunction with the immunolabeling studies (Fig. 2a), strongly suggest that the attenuation observed in response to 5-HT1A receptor activation in suicide victims is not due to a decrease in expression level of the Gi/o protein and must, therefore, result from a decrease in receptor numbers or receptor coupling.
Forskolin-stimulated and receptor-induced inhibition of adenylyl cyclase activity
We next assessed whether the decreased G protein activation detected in suicide victims also results in decreased activation of adenylyl cyclase (EC 184.108.40.206). Adenylyl cyclases belong to a family of enzymes that vary in their regulation, in particular with respect to Ca2+ levels. 5-HT1A receptor stimulation activates Gαi/o and thereby reduces the enzymatic activity of the cyclase. We therefore assayed the total activity of the cyclase (forskolin induced), as well as the inhibition induced by receptor stimulation. The activity of adenylyl cyclase in the occipital cortex of suicide victims was significantly lower at concentrations of 1, 10 and 50 µm forskolin (Fig. 3a). Moreover, receptor stimulation with 8-OH-DPAT (100 µm) in the presence of forskolin inhibited adenylyl cyclase activity less in suicide victims (Fig. 3b). These data demonstrate that both the intrinsic activity of adenylyl cyclase as well as its regulation by receptor activation are attenuated in suicide victims. The activity of adenylyl cyclase was measured in the presence or absence of EGTA (1 mm) and was found to be the same (data not shown). Therefore, the Ca2+-dependent isoforms of adenylyl cyclase are not activated by 5-HT1A receptor stimulation in the occipital cortex.
Cyclic AMP generated by adenylyl cyclase has many physiological functions (Hanoune et al. 1997). One of its major targets is PKA (EC 220.127.116.11), a serine–threonine kinase. We therefore measured the total PKA activity (in the presence of cAMP) as well as its basal activity (in the absence of cAMP). Basal activities of PKA were not significantly different in suicide victims compared with controls (data not shown), whereas the total activity showed only a non-significant trend toward lower activity in the occipital cortex of suicide victims (Fig. 4). It is interesting to note that the activity of the kinase showed an inverse U shape as 1 mm cAMP produces less stimulatory effect than lower concentrations.
Measurements of levels of non-activated and activated p44/42 MAP kinase
The 5-HT1A receptor is not only coupled to adenylyl cyclase but also to p44/42 MAPK (ERK1/2) (Cowen et al. 1996). Measurement of the protein levels of ERK1/2 by quantitative western blot analysis revealed higher levels of expression of p42 MAPK than p44 MAPK (Fig. 5a). Moreover, the levels of p44 MAPK were significantly lower in brain tissue from suicide victims (Fig. 5a). We further examined the p42/44 MAP kinase activity by western blot analysis using an antibody specific to phospho-p44 MAP kinase, which detects the total phosphorylation levels of ERK1/2. We found that the phosphorylation levels of phospho-p44 and phospho-p42 MAPK were significantly lower in suicide victims (Fig. 5b).
PI3-K (EC 18.104.22.168) is a downstream effector of 5-HT1A receptor activation in transfected human cells (Cowen et al. 1996), so we tested the activity of this enzyme in total lysate of occipital cortex. We found that the activity of PI3-K was significantly decreased in the occipital cortex of suicide victims (Figs 6a and b).
Levels and activity of endogenous protein kinase B (Akt)
Numerous studies have demonstrated that PI3-K may couple the activation of G proteins to the stimulation of the serine–threonine kinase Akt, which is involved in a variety of essential physiological responses. We measured the levels and assayed the activity of Akt in the total lysates. Whereas there was only a trend toward decreased protein levels of Akt in suicide victims (Fig. 7a), the Akt activity was significantly lower in suicide victims than in controls (Fig. 7b).
Phospholipid phosphatase PTEN
PIP3, the lipid that is generated by PI3-K, has been shown to activate phosphatidylinositol-dependent kinase (PDK-1), which in turn activates Akt (Le Good et al. 1998). The lower activity of Akt (Fig. 7) might therefore be due to decreased activity of PI3-K (Fig. 6). The reduced activity of Akt might, however, also result from hydrolysis of PIP3, the product of PI3-K, by the phospholipid phosphatase PTEN (EC 22.214.171.124). We therefore measured the levels of PTEN in the occipital cortex of controls and suicide victims. PTEN levels in suicide victims were higher (Fig. 8). Taken together, the data show that decreased activity of PI3-K, an enzyme that synthesizes PIP3, and increased levels of PTEN, an enzyme that dephosphorylates PIP3, result in decreased PDK-1 activity and therefore diminished Akt activity. A scheme of the pathways involved in 5-HT1A signaling is shown in Fig. 9.
Studies from several laboratories, including our own, have identified abnormal serotonergic transmission as part of the etiology of depression and suicide (for review see Mann 1998). The present study extends these observations to the signal cascades downstream of 5-HT1A receptor activation. In the occipital cortex of depressed suicide subjects there are significant abnormalities in major transduction pathways, including the activation of adenylyl cyclase and MAPK. The study also demonstrates, for the first time, attenuated activity of PI3-K/Akt.
One of the most direct measures of receptor activation is the increase in [35S]GTPγS binding seen upon addition of agonist. We found that the maximal effect of 8-OH-DPAT in stimulating [35S]GTPγS binding was less in suicide victims. Two previous postmortem studies have demonstrated either raised levels of [35S]GTPγS binding in frontal cortex of patients who suffered from bipolar disorder (Friedman and Wang 1996) or no change in [35S]GTPγS binding in the PFC of suicide victims (Gonzalez-Maeso et al. 2002). There are several explanations for this apparent discrepancy. First, the subjects studied by Friedman and Wang had bipolar disease and not a MDD (only one committed suicide). Second, different brain regions were examined in the two studies. Alternatively, the discrepancy may be explained by the fact that Friedman and colleagues measured the initial rate of [35S]GTPγS binding, whereas our studies were carried out at equilibrium.
Moreover, we found no changes in the levels of Gαs, Gαi/o and Gαq/11. This is in agreement with the studies of Cowburn et al. (1994) and Gonzalez-Maeso et al. (2002), who detected no changes in G protein levels in the PFC, and Dowlatshahi et al. (1999), who detected no differences in the levels of G proteins in the temporal and occipital cortex of suicide victims. In contrast, higher postmortem levels of Gαs and Gαq/11 were reported in frontal cortex and in occipital cortex respectively of subjects with bipolar disorder (Young et al. 1993; Friedman and Wang 1996; Mathews et al. 1997). Raised levels of Gαs were also reported in the PFC of depressed suicide victims (Pacheco et al. 1996). This disagreement may reflect differences in the clinical samples, treatment effects or technical factors such as the assays. We verified that the analysis was performed within the linear range, by using an internal standard, and by applying two independent techniques for the assay of Gαi/o. It is therefore concluded that the number of 5-HT1A receptors and/or the coupling efficacy to G proteins is attenuated in suicide victims. This finding is consistent with the observation that 5-HT1A receptor numbers are reduced in the occipital cortex of depressed subjects (Drevets et al. 1999; Sargent et al. 2000).
Adenylyl cyclase is a major effector of the 5-HT1A receptor. Because the coupling of the receptor to Gαi/o was reduced in suicide victims, it was expected that the ability of Gαi/o to inhibit adenylyl cyclase activity would also be attenuated. Our data suggest that both the total activity (induced by forskolin) and the inhibition of this activity (caused by stimulation of the 5-HT1A receptor by 8-OH-DPAT) are attenuated in suicide victims. Lower endogenous adenylyl cyclase activity is consistent with lower activity found in the PFC of suicide victims (Cowburn et al. 1994; Reiach et al. 1999), and in the temporal and occipital cortex (Dowlatshahi et al. 1999). These findings suggest strongly that the regulation of adenylyl cyclase is complex, and is not restricted to 5-HT receptor activation. The affected isoform(s) of adenylyl cyclase were Ca2+ independent, therefore excluding types I, III and VIII (Hanoune et al. 1997). Thus, changes in the effectors downstream of the receptor can be detected in the occipital cortex, as well as in the PFC.
Unlike Dwivedi et al. (2002), who demonstrated lower regulatory and catalytic activity of PKA in the PFC of suicide victims, we did not find such an effect in occipital cortex. Our assay, however, bypassed the measurement of regulatory subunit activity and therefore our findings reflect only the catalytic activity. The regulatory subunit activity should be measured separately in future experiments. It is interesting to note, however, that higher levels of cAMP may not only activate PKA but also activate cAMP-specific phosphodiesterases, which feed back to regulate cAMP homeostasis (for review see Conti et al. 2003).
5-HT1A receptors, in addition to regulating adenylyl cyclase, regulate MAP kinases such as the ERK1/2 kinases (Cowen et al. 1996; Della Rocca et al. 1999). Activation of the 5-HT1A receptor requires the activation and dissociation of a heterotrimeric G protein to Gα and Gβγ subunits; the two subunits can activate different transduction pathways. It is therefore possible that transduction pathways that involve adenylyl cyclase/PKA may be stimulated by Gαi/o and that transduction pathways associated with PI3-K, Akt or ERK1/2 may be stimulated by Gβγ (Guillard et al. 2003; Li et al. 2003). The hypothesis that Gβγ subunits are involved in activating PI3-K remains to be tested. Nevertheless, the activation of the 5-HT1A receptor via Gi/o is similar to the transduction pathways elucidated for the adrenergic receptor (Van Biesen et al. 1995; Touhara et al. 1995) and the Ca2+ receptor (Liu et al. 2003), in that they involve the activation of PI3-K by Gβγ subunits.
PI3-K and its downstream effector Akt are involved in cell viability, in particular of neocortical neurons, by mediating the effects of growth factors such as brain-derived neurotrophic factor (Franke et al. 1997); (Cheng et al. 2003). There is postmortem evidence of neuronal and or glial histopathology in mood disorders (Benes et al. 1998; Ongur et al. 1998; Rajkowska et al. 1999). In the case of neuronal histopathology in MDD, it was suggested that neuronal shrinkage or developmental deficiency may account for the atrophy of the neuronal soma in cortical layers. There is, however, a strong suggestion that there is also a disease-specific glial pathology in mood disorders. The glial pathology might also be related to the dysfunction of serotonergic transmission via postsynaptic receptors expressed on their soma and processes (for review see Rajkowska 2000). Most morphometric studies have been done in PFC, but one recent study reported a reduction in the volume and number of neurons in the occipital cortex of schizophrenics (Dorph-Petersen et al. 2002). Recent hypotheses of the pathology of mood disorders and antidepressant action emphasize the modulation of neoplasticity and cellular viability (Duman et al. 1997; Manji et al. 2001). Abnormalities in this pathway, including attenuated activities of PI3-K/Akt, as demonstrated in the present study, may reflect abnormal regulation during brain development.
‘Cross-talk’ between the transduction pathways depicted in Fig. 9 has been reported. At least three potential sites of convergence have been demonstrated. The first site is suggested by the observation that PI3-K activity regulates the ERK cascade (Cowen et al. 1996). A second convergence site is Akt inhibition of the ERK cascade by phosphorylation of RAF (Zimmermann and Moelling 1999). The regulation of the ERK cascade by cAMP/PKA has been studied extensively. Both positive and negative regulation has been demonstrated (for reviews see Houslay and Kolch 2000; Stork and Schmitt 2002). A third site of convergence is suggested by the observation that Akt activity is modulated by cAMP/PKA. Here too, both inhibitory (Kim et al. 2001 Lou et al. 2002) and stimulatory (Fillipa et al. 1999) effects of cAMP/PKA on Akt activity have been reported.
In the present study, however, we observed that the activities of both Akt and p44/42 MAP kinases were diminished in brains of suicide victims. Cell specificity, microdomains of signals and temporal effects may be some of the factors that dictate the direction and degree of ‘cross-talk’ between the pathways (for reviews see Houslay and Kolch 2000; Stork and Schmitt 2002). Thus, a complex regulatory mechanism, involving both positive and negative feedback, most often mediated by a member of the Ras superfamily, may affect the downstream second messengers of the 5-HT1A receptor (Fig. 9). Our findings are based on a small number of subjects tested and should be replicated in a larger study.
In the occipital cortex of depressed subjects abnormalities in 5-HT transporter (Perry et al. 1983), 5-HT1A receptor (Drevets et al. 1999; Sargent et al. 2000) and GABA levels (observed by proton magnetic resonance; Sanacora et al. 1999) indicate the involvement of this brain region in the pathophysiology of mood disorders. No such evidence exists, however, for involvement of the occipital cortex in suicidal behavior. The occipital cortex contains the primary and associative visual cortex; however, this region is also interconnected with brain regions more commonly associated with pathological emotional responses in behavior, the PFC and the limbic system. Therefore, changes in the occipital cortex might represent part of the primary pathophysiology of mood disorders, or secondary to changes in other brain regions.
In summary, we compared the activities of two apparently independent pathways downstream from 5-HT1A receptors in the occipital cortex of brain of depressed suicide victims and matched controls. We found that the coupling of this receptor to Gαi/o is decreased in suicide victims. Attenuation of coupling of this receptor to Gαi/o reduced the ligand-induced inhibition of adenylyl cyclase. The increased levels of generated cAMP, in turn, may modify the inhibition of Akt activity. The activity of Akt was also attenuated by the lower levels of PI3-K and higher levels of the lipid phosphatase PTEN. These two enzymes reciprocally regulate the levels of PIP3, which in turn activates Akt. This study supports the hypothesis that serotonergic dysfunction is involved in suicidal behavior and suggests a pivotal role for Akt in maintaining normal serotonergic transmission, perhaps involving cell resilience. Further testing this hypothesis may help in characterizing pathophysiological mechanisms in the etiology of mood disorders and suicide.
This work was supported by grants from The American Foundation for Prevention of Suicide (to HT), MH40210 (to VA) and MH62185 (to JJM).